From 2814854ee20895832c49fea08a2570606e6abe0a Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 1 Feb 2023 15:29:59 +0800 Subject: [PATCH 01/63] updated 3 articles --- TOC.md | 2 + .../information-schema-processlist.md | 85 ++++++++++--------- .../information-schema-resource-groups.md | 70 +++++++++++++++ privilege-management.md | 12 +++ 4 files changed, 130 insertions(+), 39 deletions(-) create mode 100644 information-schema/information-schema-resource-groups.md diff --git a/TOC.md b/TOC.md index dfcd7b372af7b..0afa6a8e0c3d4 100644 --- a/TOC.md +++ b/TOC.md @@ -647,6 +647,7 @@ - [`ALTER INDEX`](/sql-statements/sql-statement-alter-index.md) - [`ALTER INSTANCE`](/sql-statements/sql-statement-alter-instance.md) - [`ALTER PLACEMENT POLICY`](/sql-statements/sql-statement-alter-placement-policy.md) + - [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md) - [`ALTER TABLE`](/sql-statements/sql-statement-alter-table.md) - [`ALTER TABLE COMPACT`](/sql-statements/sql-statement-alter-table-compact.md) - [`ALTER USER`](/sql-statements/sql-statement-alter-user.md) @@ -855,6 +856,7 @@ - [`PLACEMENT_POLICIES`](/information-schema/information-schema-placement-policies.md) - [`PROCESSLIST`](/information-schema/information-schema-processlist.md) - [`REFERENTIAL_CONSTRAINTS`](/information-schema/information-schema-referential-constraints.md) + - [`RESOURCE_GROUPS`](/information-schema/information-schema-resource-groups.md) - [`SCHEMATA`](/information-schema/information-schema-schemata.md) - [`SEQUENCES`](/information-schema/information-schema-sequences.md) - [`SESSION_VARIABLES`](/information-schema/information-schema-session-variables.md) diff --git a/information-schema/information-schema-processlist.md b/information-schema/information-schema-processlist.md index 2200c5944ecb7..934448b124026 100644 --- a/information-schema/information-schema-processlist.md +++ b/information-schema/information-schema-processlist.md @@ -12,7 +12,8 @@ The `PROCESSLIST` table has additional columns not present in `SHOW PROCESSLIST` * A `DIGEST` column to show the digest of the SQL statement. * A `MEM` column to show the memory used by the request that is being processed, in bytes. * A `DISK` column to show the disk usage in bytes. -* A `TxnStart` column to show the start time of the transaction +* A `TxnStart` column to show the start time of the transaction. +* A `RESOURCE_GROUP_NAME` column to show the resource group name. {{< copyable "sql" >}} @@ -21,24 +22,25 @@ USE information_schema; DESC processlist; ``` -``` -+----------+---------------------+------+------+---------+-------+ -| Field | Type | Null | Key | Default | Extra | -+----------+---------------------+------+------+---------+-------+ -| ID | bigint(21) unsigned | NO | | 0 | | -| USER | varchar(16) | NO | | | | -| HOST | varchar(64) | NO | | | | -| DB | varchar(64) | YES | | NULL | | -| COMMAND | varchar(16) | NO | | | | -| TIME | int(7) | NO | | 0 | | -| STATE | varchar(7) | YES | | NULL | | -| INFO | longtext | YES | | NULL | | -| DIGEST | varchar(64) | YES | | | | -| MEM | bigint(21) unsigned | YES | | NULL | | -| DISK | bigint(21) unsigned | YES | | NULL | | -| TxnStart | varchar(64) | NO | | | | -+----------+---------------------+------+------+---------+-------+ -12 rows in set (0.00 sec) +```sql ++---------------------+---------------------+------+------+---------+-------+ +| Field | Type | Null | Key | Default | Extra | ++---------------------+---------------------+------+------+---------+-------+ +| ID | bigint(21) unsigned | NO | | 0 | | +| USER | varchar(16) | NO | | | | +| HOST | varchar(64) | NO | | | | +| DB | varchar(64) | YES | | NULL | | +| COMMAND | varchar(16) | NO | | | | +| TIME | int(7) | NO | | 0 | | +| STATE | varchar(7) | YES | | NULL | | +| INFO | longtext | YES | | NULL | | +| DIGEST | varchar(64) | YES | | | | +| MEM | bigint(21) unsigned | YES | | NULL | | +| DISK | bigint(21) unsigned | YES | | NULL | | +| TxnStart | varchar(64) | NO | | | | +| RESOURCE_GROUP_NAME | varchar(32) | NO | | | | ++---------------------+---------------------+------+------+---------+-------+ +13 rows in set (0.00 sec) ``` {{< copyable "sql" >}} @@ -47,18 +49,21 @@ DESC processlist; SELECT * FROM processlist\G ``` -``` +```sql *************************** 1. row *************************** - ID: 16 - USER: root - HOST: 127.0.0.1 - DB: information_schema - COMMAND: Query - TIME: 0 - STATE: autocommit - INFO: SELECT * FROM processlist - MEM: 0 -TxnStart: + ID: 5794501045170209171 + USER: root + HOST: 127.0.0.1:60645 + DB: information_schema + COMMAND: Query + TIME: 0 + STATE: autocommit + INFO: SELECT * FROM processlist + DIGEST: dbfaa16980ec628011029f0aaf0d160f4b040885240dfc567bf760d96d374f7e + MEM: 0 + DISK: 0 + TxnStart: +RESOURCE_GROUP_NAME: default 1 row in set (0.00 sec) ``` @@ -76,6 +81,7 @@ Fields in the `PROCESSLIST` table are described as follows: * MEM: The memory used by the request that is being processed, in bytes. * DISK: The disk usage in bytes. * TxnStart: The start time of the transaction. +* RESOURCE_GROUP_NAME: The resource group name. ## CLUSTER_PROCESSLIST @@ -88,13 +94,14 @@ SELECT * FROM information_schema.cluster_processlist; ``` ```sql -+-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+ -| INSTANCE | ID | USER | HOST | DB | COMMAND | TIME | STATE | INFO | MEM | TxnStart | -+-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+ -| 10.0.1.22:10080 | 150 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077223) | -| 10.0.1.22:10080 | 138 | root | 10.0.1.1 | test | Query | 0 | autocommit | SELECT * FROM information_schema.cluster_processlist | 0 | 05-28 03:54:21.230(416976223923077220) | -| 10.0.1.22:10080 | 151 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077224) | -| 10.0.1.21:10080 | 15 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077222) | -| 10.0.1.21:10080 | 14 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077225) | -+-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+ ++-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+---------------------+ +| INSTANCE | ID | USER | HOST | DB | COMMAND | TIME | STATE | INFO | MEM | TxnStart | RESOURCE_GROUP_NAME | ++-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+---------------------+ + +| 10.0.1.22:10080 | 150 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077223) | default | +| 10.0.1.22:10080 | 138 | root | 10.0.1.1 | test | Query | 0 | autocommit | SELECT * FROM information_schema.cluster_processlist | 0 | 05-28 03:54:21.230(416976223923077220) | rg1 | +| 10.0.1.22:10080 | 151 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077224) | rg2 | +| 10.0.1.21:10080 | 15 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077222) | default | +| 10.0.1.21:10080 | 14 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077225) | default | ++-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+---------------------+ ``` diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md new file mode 100644 index 0000000000000..9cc11c6a33a84 --- /dev/null +++ b/information-schema/information-schema-resource-groups.md @@ -0,0 +1,70 @@ +--- +title: RESOURCE_GROUPS +summary: Learn the `RESOURCE_GROUPS` information_schema table. +--- + +The `RESOURCE_GROUPS` table shows information about all resource groups. See [Resource Conctrol in TiDB](/tidb-resource-control.md). + +```sql +USE information_schema; +DESC resource_groups; +``` + +```sql ++-----------------+-------------+------+------+---------+-------+ +| Field | Type | Null | Key | Default | Extra | ++-----------------+-------------+------+------+---------+-------+ +| NAME | varchar(32) | NO | | NULL | | +| MODE | varchar(12) | NO | | NULL | | +| RRU | bigint(21) | YES | | NULL | | +| RRU_TOKEN | bigint(21) | YES | | NULL | | +| WRU | bigint(21) | YES | | NULL | | +| WRU_TOKEN | bigint(21) | YES | | NULL | | +| CPU | bigint(21) | YES | | NULL | | +| READ_BANDWIDTH | bigint(21) | YES | | NULL | | +| WRITE_BANDWIDTH | bigint(21) | YES | | NULL | | ++-----------------+-------------+------+------+---------+-------+ +9 rows in set (0.08 sec) +``` + +## Examples + +{{< copyable "sql" >}} + +```sql +mysql> CREATE RESOURCE GROUP rg1 RRU_PER_SEC=1000 WRU_PER_SEC=2000 ; -- Use RU_MODE to create the resource group rg1 +Query OK, 0 rows affected (0.34 sec) +mysql> SHOW CREATE RESOURCE GROUP rg1; -- Displays the definition of the rg1 resource group. ++----------------+------------------------------------------------------------------+ +| Resource_Group | Create Resource Group | ++----------------+------------------------------------------------------------------+ +| rg1 | CREATE RESOURCE GROUP `rg1` RRU_PER_SEC=1000 WRU_PER_SEC=2000 | ++----------------+------------------------------------------------------------------+ +1 row in set (0.00 sec) +mysql> SELECT * FROM information_schema.placement_policies WHERE NAME = 'rg1'; -- Displays the runtime status of the resource group rg1. ++--------+---------+------+-----------+------+-----------+------+----------------+-----------------+ +| NAME | MODE | RRU | RRU_TOKEN | WRU | WRU_TOKEN | CPU | READ_BANDWIDTH | WRITE_BANDWIDTH | ++--------+---------+------+-----------+------+-----------+------+----------------+-----------------+ +| rg1 | RU_MODE | 1000 | 168999 | 2000 | 249999 | NULL | NULL | NULL | ++--------+---------+------+-----------+------+-----------+------+----------------+-----------------+ +1 row in set (0.02 sec) +``` + +```sql +mysql> CREATE RESOURCE GROUP rg2 CPU="8" IO_READ_BANDWIDTH="1000Mi" IO_WRITE_BANDWIDTH="1000Mi"; -- Use RAW_MODE to create the resource group rg2 +Query OK, 0 rows affected (0.20 sec) +mysql> show create resource group rg2; -- Displays the definition of the rg2 resource group. ++----------------+-----------------------------------------------------------------------------------------------+ +| Resource_Group | Create Resource Group | ++----------------+-----------------------------------------------------------------------------------------------+ +| rg2 | CREATE RESOURCE GROUP `rg2` CPU="8" IO_READ_BANDWIDTH="1000Mi" IO_WRITE_BANDWIDTH="1000Mi" | ++----------------+-----------------------------------------------------------------------------------------------+ +1 row in set (0.00 sec) +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg2'; -- Displays the runtime status of the resource group rg2. ++--------+----------+------+-----------+------+-----------+------+----------------+-----------------+ +| NAME | MODE | RRU | RRU_TOKEN | WRU | WRU_TOKEN | CPU | READ_BANDWIDTH | WRITE_BANDWIDTH | ++--------+----------+------+-----------+------+-----------+------+----------------+-----------------+ +| rg2 | RAW_MODE | NULL | NULL | NULL | NULL | 8000 | 1048576000 | 1048576000 | ++--------+----------+------+-----------+------+-----------+------+----------------+-----------------+ +1 row in set (0.00 sec) +``` diff --git a/privilege-management.md b/privilege-management.md index c38bed48a207d..118844b878ac7 100644 --- a/privilege-management.md +++ b/privilege-management.md @@ -363,6 +363,18 @@ Requires `SUPER` privilege. Requires `SUPER` or `CONNECTION_ADMIN` privilege to kill other user sessions. +### CREATE RESOURCE GROUP + +Requires `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. + +### ALTER RESOURCE GROUP + +Requires `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. + +### DROP RESOURCE GROUP + +Requires `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. + ## Implementation of the privilege system ### Privilege table From 1197f87281784f22985faa426f201e010068769f Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 1 Feb 2023 15:52:24 +0800 Subject: [PATCH 02/63] Create sql-statement-alter-resource-group.md --- .../sql-statement-alter-resource-group.md | 92 +++++++++++++++++++ 1 file changed, 92 insertions(+) create mode 100644 sql-statements/sql-statement-alter-resource-group.md diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md new file mode 100644 index 0000000000000..4ea13893c7c43 --- /dev/null +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -0,0 +1,92 @@ +--- +title: ALTER RESOURCE GROUP +summary: Learn the usage of ALTER RESOURCE GROUP in TiDB. +--- + +# ALTER RESOURCE GROUP + +The `ALTER RESOURCE GROUP` statement is used to modify the resource group in a database. + +## Synopsis + +```ebnf+diagram +AlterResourceGroupStmt: + "ALTER" "RESOURCE" "GROUP" IfNotExists ResourceGroupName ResourceGroupOptionList BurstableOption + +IfNotExists ::= + ('IF' 'NOT' 'EXISTS')? + +ResourceGroupName: + Identifier + +ResourceGroupOptionList: + DirectResourceGroupOption +| ResourceGroupOptionList DirectResourceGroupOption +| ResourceGroupOptionList ',' DirectResourceGroupOption + +DirectResourceGroupOption: + "RRU_PER_SEC" EqOpt stringLit +| "WRU_PER_SEC" EqOpt stringLit + +BurstableOption ::= + ("BURSTABLE")? + +``` + +TiDB supports the following `DirectResourceGroupOption`, where [RU (Resource Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. + +| Parameter | Description | Example | +|---------------|-------------------------------------|------------------------| +|`RRU_PER_SEC` | Quota of RU read per second |`RRU_PER_SEC = 500` | +|`WRU_PER_SEC` | Quota of RU write per second |`WRU_PER_SEC = 300` | + +If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota if the system resources are sufficient. + +> **Note:** +> +> The `ALTER RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_group`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. + +## Examples + +Create a resource group named `rg1` and modify its properties. + +```sql +mysql> DROP RESOURCE GROUP IF EXISTS rg1; +Query OK, 0 rows affected (0.22 sec) +mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 + -> RRU_PER_SEC = 500 + -> WRU_PER_SEC = 300 + -> BURSTABLE + -> ; +Query OK, 0 rows affected (0.08 sec) +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; ++------+--------------+---------------------------------------------------------------+ +| Name | Plan_type | Directive | ++------+--------------+---------------------------------------------------------------+ +| rg1 | tenancy | {"RRU_PER_SEC": 500, "WRU_PER_SEC": 300, "BURSTABLE": true} | ++------+--------------+---------------------------------------------------------------+ +1 row in set (0.00 sec) + +mysql> ALTER RESOURCE GROUP IF NOT EXISTS rg1 + -> RRU_PER_SEC = 600 + -> WRU_PER_SEC = 400 + -> ; +Query OK, 0 rows affected (0.09 sec) +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; ++------+--------------+---------------------------------------------------------------+ +| Name | Plan_type | Directive | ++------+--------------+---------------------------------------------------------------+ +| rg1 | tenancy | {"RRU_PER_SEC": 600, "WRU_PER_SEC": 400, "BURSTABLE": false} | ++------+--------------+---------------------------------------------------------------+ +1 row in set (0.00 sec) +``` + +## MySQL compatibility + +MySQL also supports [ALTER RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/en/alter-resource-group.html). However, the accepted parameters are different from TiDB so that they are not compatible. + +## See also + +* [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) +* [CREATE RESOURCE GROUP](/sql-statements/sql-statement-create-resource-group.md) +* [RU](/tidb-RU.md) From 7aa88984ef4cedfd443bc5ded74ee88a65d235db Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 1 Feb 2023 18:11:05 +0800 Subject: [PATCH 03/63] Update sql-statement-alter-user.md --- sql-statements/sql-statement-alter-user.md | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/sql-statements/sql-statement-alter-user.md b/sql-statements/sql-statement-alter-user.md index 0626a5eb93d19..f93c7c457a552 100644 --- a/sql-statements/sql-statement-alter-user.md +++ b/sql-statements/sql-statement-alter-user.md @@ -11,7 +11,7 @@ This statement changes an existing user inside the TiDB privilege system. In the ```ebnf+diagram AlterUserStmt ::= - 'ALTER' 'USER' IfExists (UserSpecList RequireClauseOpt ConnectionOptions PasswordOption LockOption AttributeOption | 'USER' '(' ')' 'IDENTIFIED' 'BY' AuthString) + 'ALTER' 'USER' IfExists (UserSpecList RequireClauseOpt ConnectionOptions PasswordOption LockOption AttributeOption | 'USER' '(' ')' 'IDENTIFIED' 'BY' AuthString) ResourceGroupNameOption UserSpecList ::= UserSpec ( ',' UserSpec )* @@ -30,6 +30,8 @@ PasswordOption ::= ( 'PASSWORD' 'EXPIRE' ( 'DEFAULT' | 'NEVER' | 'INTERVAL' N 'D LockOption ::= ( 'ACCOUNT' 'LOCK' | 'ACCOUNT' 'UNLOCK' )? AttributeOption ::= ( 'COMMENT' CommentString | 'ATTRIBUTE' AttributeString )? + +ResourceGroupNameOption::= ( 'RESOURCE' 'GROUP' Identifier)? ``` ## Examples @@ -134,6 +136,16 @@ ALTER USER 'newuser' PASSWORD REUSE INTERVAL 90 DAY; Query OK, 0 rows affected (0.02 sec) ``` +Use `ALTER USER ... RESOURCE GROUP` to modify the resource group of the user `newuser` to `rg1`. + +```sql +ALTER USER 'newuser' RESOURCE GROUP rg1; +``` + +``` +Query OK, 0 rows affected (0.02 sec) +``` + ## See also From 0ff80b28fd417bea3fdf44e9618b953721aeb918 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 1 Feb 2023 18:20:14 +0800 Subject: [PATCH 04/63] added CREATE RESOURCE GROUP --- TOC.md | 1 + .../sql-statement-create-resource-group.md | 81 +++++++++++++++++++ 2 files changed, 82 insertions(+) create mode 100644 sql-statements/sql-statement-create-resource-group.md diff --git a/TOC.md b/TOC.md index 0afa6a8e0c3d4..40cc3e5c97acc 100644 --- a/TOC.md +++ b/TOC.md @@ -663,6 +663,7 @@ - [`CREATE DATABASE`](/sql-statements/sql-statement-create-database.md) - [`CREATE INDEX`](/sql-statements/sql-statement-create-index.md) - [`CREATE PLACEMENT POLICY`](/sql-statements/sql-statement-create-placement-policy.md) + - [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md) - [`CREATE ROLE`](/sql-statements/sql-statement-create-role.md) - [`CREATE SEQUENCE`](/sql-statements/sql-statement-create-sequence.md) - [`CREATE TABLE LIKE`](/sql-statements/sql-statement-create-table-like.md) diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md new file mode 100644 index 0000000000000..f6e1c9df64bf3 --- /dev/null +++ b/sql-statements/sql-statement-create-resource-group.md @@ -0,0 +1,81 @@ +--- +title: CREATE RESOURCE GROUP +summary: Learn the usage of CREATE RESOURCE GROUP in TiDB. +--- + +# CREATE RESOURCE GROUP + +You can run the `CREATE RESOURCE GROUP` statement to create a resource group in the currently selected database. + +## Synopsis + +```ebnf+diagram +CreateResourceGroupStmt: + "CREATE" "RESOURCE" "GROUP" IfNotExists ResourceGroupName ResourceGroupOptionList BurstableOption + +IfNotExists ::= + ('IF' 'NOT' 'EXISTS')? + +ResourceGroupName: + Identifier + +ResourceGroupOptionList: + DirectResourceGroupOption +| ResourceGroupOptionList DirectResourceGroupOption +| ResourceGroupOptionList ',' DirectResourceGroupOption + +DirectResourceGroupOption: + "RRU_PER_SEC" EqOpt stringLit +| "WRU_PER_SEC" EqOpt stringLit + +BurstableOption ::= + ("BURSTABLE")? + +``` + +The `ResourceGroupName` of a resource group is globally unique and cannot be duplicated. + +TiDB supports the following `DirectResourceGroupOption`, where [RU (Resource Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. + +| Parameter | Description | Example | +|---------------|-------------------------------------|------------------------| +|`RRU_PER_SEC` | Quota of RU read per second |`RRU_PER_SEC = 500` | +|`WRU_PER_SEC` | Quota of RU write per second |`WRU_PER_SEC = 300` | + +If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota if the system resources are sufficient. + +> **Note:** +> +> The `CREATE RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_group`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. + +## Examples + +Create a resource group named `rg1`. + +```sql +mysql> DROP RESOURCE GROUP IF EXISTS rg1; +Query OK, 0 rows affected (0.22 sec) +mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 + -> RRU_PER_SEC = 500 + -> WRU_PER_SEC = 300 + -> BURSTABLE + -> ; +Query OK, 0 rows affected (0.08 sec) +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; ++------+--------------+---------------------------------------------------------------+ +| Name | Plan_type | Directive | ++------+--------------+---------------------------------------------------------------+ +| rg1 | tenancy | {"RRU_PER_SEC": 5000, "WRU_PER_SEC": 1000, "BURSTABLE": true} | ++------+--------------+---------------------------------------------------------------+ +1 row in set (0.00 sec) +``` + +## MySQL compatibility + +MySQL also supports [CREATE RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/en/create-resource-group.html). However, the accepted parameters are different from TiDB so that they are not compatible. + +## See also + +* [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) +* [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) +* [RU](/tidb-RU.md) From 30d94860a2c2371e2d1a6454e9511e52431703f0 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 1 Feb 2023 18:45:14 +0800 Subject: [PATCH 05/63] translate 6 more docs --- TOC.md | 2 + .../sql-statement-create-resource-group.md | 2 +- sql-statements/sql-statement-create-user.md | 20 +++- .../sql-statement-drop-resource-group.md | 66 +++++++++++++ ...ql-statement-show-create-resource-group.md | 46 +++++++++ .../sql-statement-show-privileges.md | 93 +++++++++++-------- system-variables.md | 13 +++ 7 files changed, 202 insertions(+), 40 deletions(-) create mode 100644 sql-statements/sql-statement-drop-resource-group.md create mode 100644 sql-statements/sql-statement-show-create-resource-group.md diff --git a/TOC.md b/TOC.md index 40cc3e5c97acc..d834e1fcb0d8d 100644 --- a/TOC.md +++ b/TOC.md @@ -680,6 +680,7 @@ - [`DROP DATABASE`](/sql-statements/sql-statement-drop-database.md) - [`DROP INDEX`](/sql-statements/sql-statement-drop-index.md) - [`DROP PLACEMENT POLICY`](/sql-statements/sql-statement-drop-placement-policy.md) + - [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md) - [`DROP ROLE`](/sql-statements/sql-statement-drop-role.md) - [`DROP SEQUENCE`](/sql-statements/sql-statement-drop-sequence.md) - [`DROP STATS`](/sql-statements/sql-statement-drop-stats.md) @@ -731,6 +732,7 @@ - [`SHOW CONFIG`](/sql-statements/sql-statement-show-config.md) - [`SHOW CREATE DATABASE`](/sql-statements/sql-statement-show-create-database.md) - [`SHOW CREATE PLACEMENT POLICY`](/sql-statements/sql-statement-show-create-placement-policy.md) + - [`SHOW CREATE RESOURCE GROUP`](/sql-statements/sql-statement-show-create-resource-group.md) - [`SHOW CREATE SEQUENCE`](/sql-statements/sql-statement-show-create-sequence.md) - [`SHOW CREATE TABLE`](/sql-statements/sql-statement-show-create-table.md) - [`SHOW CREATE USER`](/sql-statements/sql-statement-show-create-user.md) diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index f6e1c9df64bf3..e9bb21b3910ff 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -5,7 +5,7 @@ summary: Learn the usage of CREATE RESOURCE GROUP in TiDB. # CREATE RESOURCE GROUP -You can run the `CREATE RESOURCE GROUP` statement to create a resource group in the currently selected database. +You can use the `CREATE RESOURCE GROUP` statement to create a resource group in the currently selected database. ## Synopsis diff --git a/sql-statements/sql-statement-create-user.md b/sql-statements/sql-statement-create-user.md index f73221ff90b63..607240c59957c 100644 --- a/sql-statements/sql-statement-create-user.md +++ b/sql-statements/sql-statement-create-user.md @@ -11,7 +11,7 @@ This statement creates a new user, specified with a password. In the MySQL privi ```ebnf+diagram CreateUserStmt ::= - 'CREATE' 'USER' IfNotExists UserSpecList RequireClauseOpt ConnectionOptions PasswordOption LockOption AttributeOption + 'CREATE' 'USER' IfNotExists UserSpecList RequireClauseOpt ConnectionOptions PasswordOption LockOption AttributeOption ResourceGroupNameOption IfNotExists ::= ('IF' 'NOT' 'EXISTS')? @@ -34,6 +34,8 @@ PasswordOption ::= ( 'PASSWORD' 'EXPIRE' ( 'DEFAULT' | 'NEVER' | 'INTERVAL' N 'D LockOption ::= ( 'ACCOUNT' 'LOCK' | 'ACCOUNT' 'UNLOCK' )? AttributeOption ::= ( 'COMMENT' CommentString | 'ATTRIBUTE' AttributeString )? + +ResourceGroupNameOption::= ( 'RESOURCE' 'GROUP' Identifier)? ``` ## Examples @@ -128,6 +130,22 @@ CREATE USER 'newuser9'@'%' PASSWORD EXPIRE; Query OK, 1 row affected (0.02 sec) ``` +Create a user that uses the resource group `rg1`. + +```sql +CREATE USER 'newuser7'@'%' RESOURCE GROUP rg1; +SELECT USER, HOST, USER_ATTRIBUTES FROM MYSQL.USER WHERE USER='newuser7'; +``` + +```sql ++-----------+------+---------------------------------------------------+ +| USER | HOST | ATTRIBUTE | ++-----------+------+---------------------------------------------------+ +| newuser7 | % | {"resource_group": "rg1"} | ++-----------+------+---------------------------------------------------+ +1 rows in set (0.00 sec) +``` + ## MySQL compatibility The following `CREATE USER` options are not yet supported by TiDB, and will be parsed but ignored: diff --git a/sql-statements/sql-statement-drop-resource-group.md b/sql-statements/sql-statement-drop-resource-group.md new file mode 100644 index 0000000000000..9c68334564084 --- /dev/null +++ b/sql-statements/sql-statement-drop-resource-group.md @@ -0,0 +1,66 @@ +--- +title: DROP RESOURCE GROUP +summary: Learn the usage of DROP RESOURCE GROUP in TiDB. +--- + +# DROP RESOURCE GROUP + +You can use the `DROP RESOURCE GROUP` statement to drop a resource group in the currently selected database. + +## Synopsis + +```ebnf+diagram +DropResourceGroupStmt: + "DROP" "RESOURCE" "GROUP" IfNotExists ResourceGroupName + +IfNotExists ::= + ('IF' 'NOT' 'EXISTS')? + +ResourceGroupName: + Identifier +``` + +> **Note:** +> +> The `DROP RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_group`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. + +## Examples + +Drop a resource group named `rg1`. + +```sql +mysql> DROP RESOURCE GROUP IF EXISTS rg1; +Query OK, 0 rows affected (0.22 sec) +mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 ( + -> RRU_PER_SEC = 500 + -> WRU_PER_SEC = 300 + -> BURSTABLE + -> ); +Query OK, 0 rows affected (0.08 sec) +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; ++------+--------------+---------------------------------------------------------------+ +| Name | Plan_type | Directive | ++------+--------------+---------------------------------------------------------------+ +| rg1 | tenancy | {"RRU_PER_SEC": 500, "WRU_PER_SEC": 300, "BURSTABLE": true} | ++------+--------------+---------------------------------------------------------------+ +1 row in set (0.00 sec) + +mysql> DROP RESOURCE GROUP IF NOT EXISTS rg1 ; +Query OK, 1 rows affected (0.09 sec) +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; ++------+--------------+---------------------------------------------------------------+ +| Name | Plan_type | Directive | ++------+--------------+---------------------------------------------------------------+ ++------+--------------+---------------------------------------------------------------+ +0 row in set (0.00 sec) +``` + +## MySQL compatibility + +MySQL also supports [DROP RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/en/drop-resource-group.html), but TiDB does not support the `FORCE` parameter. + +## See also + +* [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) +* [CREATE RESOURCE GROUP](/sql-statements/sql-statement-create-resource-group.md) +* [RU](/tidb-RU.md) \ No newline at end of file diff --git a/sql-statements/sql-statement-show-create-resource-group.md b/sql-statements/sql-statement-show-create-resource-group.md new file mode 100644 index 0000000000000..feacae4360619 --- /dev/null +++ b/sql-statements/sql-statement-show-create-resource-group.md @@ -0,0 +1,46 @@ +--- +title: SHOW CREATE RESOURCE GROUP +summary: Learn the usage of SHOW CREATE RESOURCE GROUP in TiDB. +--- + +# SHOW CREATE RESOURCE GROUP + +You can use the `SHOW CREATE RESOURCE GROUP` statement to view the current definition of a resource group, and recreate it in another TiDB cluster. + +## Synopsis + +```ebnf+diagram +ShowCreateResourceGroupStmt ::= + "SHOW" "CREATE" "RESOURCE" "GROUP" ResourceGroupName + +ResourceGroupName ::= + Identifier +``` + +## Examples + +View the current definition of the resource group `rg1` and recreate it in another TiDB cluster. + +```sql +CREATE RESOURCE GROUP rg1 RRU_PER_SEC=100 WRU_PER_SEC=200; +Query OK, 0 rows affected (0.10 sec) +``` + +```sql +SHOW CREATE RESOURCE GROUP rg1; +***************************[ 1. row ]*************************** +Resource_Group | rg1 +Create Resource Group | CREATE RESOURCE GROUP `rg1` RRU_PER_SEC=100 WRU_PER_SEC=200 +1 row in set (0.00 sec) +``` + +## MySQL compatibility + +This statement is a TiDB extension for MySQL. + +## See also + +* [TiDB RESOURCE CONTROL](/tidb-resource-control.md) +* [CREATE RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) +* [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) +* [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) diff --git a/sql-statements/sql-statement-show-privileges.md b/sql-statements/sql-statement-show-privileges.md index c5b4999ac6396..8b3852aaeb607 100644 --- a/sql-statements/sql-statement-show-privileges.md +++ b/sql-statements/sql-statement-show-privileges.md @@ -15,45 +15,62 @@ This statement shows a list of assignable privileges in TiDB. It is a static lis ## Examples -```sql +``` mysql> show privileges; -+-------------------------+---------------------------------------+-------------------------------------------------------+ -| Privilege | Context | Comment | -+-------------------------+---------------------------------------+-------------------------------------------------------+ -| Alter | Tables | To alter the table | -| Alter | Tables | To alter the table | -| Alter routine | Functions,Procedures | To alter or drop stored functions/procedures | -| Create | Databases,Tables,Indexes | To create new databases and tables | -| Create routine | Databases | To use CREATE FUNCTION/PROCEDURE | -| Create temporary tables | Databases | To use CREATE TEMPORARY TABLE | -| Create view | Tables | To create new views | -| Create user | Server Admin | To create new users | -| Delete | Tables | To delete existing rows | -| Drop | Databases,Tables | To drop databases, tables, and views | -| Event | Server Admin | To create, alter, drop and execute events | -| Execute | Functions,Procedures | To execute stored routines | -| File | File access on server | To read and write files on the server | -| Grant option | Databases,Tables,Functions,Procedures | To give to other users those privileges you possess | -| Index | Tables | To create or drop indexes | -| Insert | Tables | To insert data into tables | -| Lock tables | Databases | To use LOCK TABLES (together with SELECT privilege) | -| Process | Server Admin | To view the plain text of currently executing queries | -| Proxy | Server Admin | To make proxy user possible | -| References | Databases,Tables | To have references on tables | -| Reload | Server Admin | To reload or refresh tables, logs and privileges | -| Replication client | Server Admin | To ask where the slave or master servers are | -| Replication slave | Server Admin | To read binary log events from the master | -| Select | Tables | To retrieve rows from table | -| Show databases | Server Admin | To see all databases with SHOW DATABASES | -| Show view | Tables | To see views with SHOW CREATE VIEW | -| Shutdown | Server Admin | To shut down the server | -| Super | Server Admin | To use KILL thread, SET GLOBAL, CHANGE MASTER, etc. | -| Trigger | Tables | To use triggers | -| Create tablespace | Server Admin | To create/alter/drop tablespaces | -| Update | Tables | To update existing rows | -| Usage | Server Admin | No privileges - allow connect only | -+-------------------------+---------------------------------------+-------------------------------------------------------+ -32 rows in set (0.00 sec) ++---------------------------------+---------------------------------------+-------------------------------------------------------+ +| Privilege | Context | Comment | ++---------------------------------+---------------------------------------+-------------------------------------------------------+ +| Alter | Tables | To alter the table | +| Alter routine | Functions,Procedures | To alter or drop stored functions/procedures | +| Config | Server Admin | To use SHOW CONFIG and SET CONFIG statements | +| Create | Databases,Tables,Indexes | To create new databases and tables | +| Create routine | Databases | To use CREATE FUNCTION/PROCEDURE | +| Create role | Server Admin | To create new roles | +| Create temporary tables | Databases | To use CREATE TEMPORARY TABLE | +| Create view | Tables | To create new views | +| Create user | Server Admin | To create new users | +| Delete | Tables | To delete existing rows | +| Drop | Databases,Tables | To drop databases, tables, and views | +| Drop role | Server Admin | To drop roles | +| Event | Server Admin | To create, alter, drop and execute events | +| Execute | Functions,Procedures | To execute stored routines | +| File | File access on server | To read and write files on the server | +| Grant option | Databases,Tables,Functions,Procedures | To give to other users those privileges you possess | +| Index | Tables | To create or drop indexes | +| Insert | Tables | To insert data into tables | +| Lock tables | Databases | To use LOCK TABLES (together with SELECT privilege) | +| Process | Server Admin | To view the plain text of currently executing queries | +| Proxy | Server Admin | To make proxy user possible | +| References | Databases,Tables | To have references on tables | +| Reload | Server Admin | To reload or refresh tables, logs and privileges | +| Replication client | Server Admin | To ask where the slave or master servers are | +| Replication slave | Server Admin | To read binary log events from the master | +| Select | Tables | To retrieve rows from table | +| Show databases | Server Admin | To see all databases with SHOW DATABASES | +| Show view | Tables | To see views with SHOW CREATE VIEW | +| Shutdown | Server Admin | To shut down the server | +| Super | Server Admin | To use KILL thread, SET GLOBAL, CHANGE MASTER, etc. | +| Trigger | Tables | To use triggers | +| Create tablespace | Server Admin | To create/alter/drop tablespaces | +| Update | Tables | To update existing rows | +| Usage | Server Admin | No privileges - allow connect only | +| BACKUP_ADMIN | Server Admin | | +| RESTORE_ADMIN | Server Admin | | +| SYSTEM_USER | Server Admin | | +| SYSTEM_VARIABLES_ADMIN | Server Admin | | +| ROLE_ADMIN | Server Admin | | +| CONNECTION_ADMIN | Server Admin | | +| PLACEMENT_ADMIN | Server Admin | | +| DASHBOARD_CLIENT | Server Admin | | +| RESTRICTED_TABLES_ADMIN | Server Admin | | +| RESTRICTED_STATUS_ADMIN | Server Admin | | +| RESTRICTED_VARIABLES_ADMIN | Server Admin | | +| RESTRICTED_USER_ADMIN | Server Admin | | +| RESTRICTED_CONNECTION_ADMIN | Server Admin | | +| RESTRICTED_REPLICA_WRITER_ADMIN | Server Admin | | +| RESOURCE_GROUP_ADMIN | Server Admin | | ++---------------------------------+---------------------------------------+-------------------------------------------------------+ +49 rows in set (0.00 sec) ``` ## MySQL compatibility diff --git a/system-variables.md b/system-variables.md index be8a082f557cd..030ada0d437f7 100644 --- a/system-variables.md +++ b/system-variables.md @@ -1824,6 +1824,19 @@ MPP is a distributed computing framework provided by the TiFlash engine, which a +### `tidb_enable_resource_control` New in v6.6.0 + +> **Warning:** +> +> This feature is experimental and its form and usage may change in subsequent versions. + +- Scope: GLOBAL +- Persists to cluster: Yes +- Default value: `OFF` +- Type: Boolean +- This variable is a switch for [the resource control feature](/tidb-resource-control.md). When this variable is set to `ON`, the cluster supports resource isolation of applications by resource group. + + ### tidb_enable_reuse_chunk New in v6.4.0 - Scope: SESSION | GLOBAL From 97b64e7a3e584b9d51085925767c3297560cce25 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 1 Feb 2023 21:53:51 +0800 Subject: [PATCH 06/63] add part of resource control doc --- TOC.md | 1 + tidb-resource-control.md | 38 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 39 insertions(+) create mode 100644 tidb-resource-control.md diff --git a/TOC.md b/TOC.md index d834e1fcb0d8d..97d2d7dbf80f7 100644 --- a/TOC.md +++ b/TOC.md @@ -159,6 +159,7 @@ - [Back up and Restore Data Using Dumpling and TiDB Lightning](/backup-and-restore-using-dumpling-lightning.md) - [Back Up and Restore RawKV](/br/rawkv-backup-and-restore.md) - [Incremental Backup and Restore](/br/br-incremental-guide.md) + - [Resource Control](/tidb-resource-control.md) - [Configure Time Zone](/configure-time-zone.md) - [Daily Checklist](/daily-check.md) - [Maintain TiFlash](/tiflash/maintain-tiflash.md) diff --git a/tidb-resource-control.md b/tidb-resource-control.md new file mode 100644 index 0000000000000..a2bb9db5bc156 --- /dev/null +++ b/tidb-resource-control.md @@ -0,0 +1,38 @@ +--- +title: Use Resource Control to Achieve Resource Isolation +summary: Learn how to use resource management to control and schedule application resource consumption. +--- + +# Use Resource Control to Achieve Resource Isolation + +> **Warning:** +> +> This feature is experimental and its form and usage may change in subsequent versions. + +Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas by resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. + +The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. + +This feature applies to the following scenarios: + +- You can combine multiple small and medium-sized applications from different systems into one TiDB cluster. If the load of an individual application grows larger, it does not affect the normal operation of other businesses. When the system load is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. +- You can choose to combine all test environments into a single cluster, or group the batch tasks that consume more resources into a single resource group. It can improve hardware utilization and reduce operating costs while ensuring that critical applications receive the necessary resources. + +In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. + +## Parameters for resource control + +The resource control feature introduces two new global variables. + +* TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control) to control whether to enable flow control for a resource group. +* TiKV: you can use the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. Note that this parameter does not support dynamic modification. You need to restart the TiKV instance to for the modification to take effect. + +The results of the combination of these two parameters are shown in the following table. + +| `resource_control.enabled` | `tidb_enable_resource_control`= ON | `tidb_enable_resource_control`= OFF | +|:----------------------------|:-------------------------------------|:-------------------------------------| +| `resource_control.enabled`= true | Flow control and scheduling (recommended) | Invalid combination | +| `resource_control.enabled`= false | Only flow control | The feature is disabled. | + +## How to use resource control + From 83712f5bdc6d8558ce91b2bdc32db124e5af46e5 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 2 Feb 2023 10:44:57 +0800 Subject: [PATCH 07/63] finish translation except ru.md --- ...ql-statement-show-create-resource-group.md | 2 +- tidb-resource-control.md | 102 +++++++++++++++++- tikv-configuration-file.md | 9 ++ 3 files changed, 111 insertions(+), 2 deletions(-) diff --git a/sql-statements/sql-statement-show-create-resource-group.md b/sql-statements/sql-statement-show-create-resource-group.md index feacae4360619..a0e293a481468 100644 --- a/sql-statements/sql-statement-show-create-resource-group.md +++ b/sql-statements/sql-statement-show-create-resource-group.md @@ -5,7 +5,7 @@ summary: Learn the usage of SHOW CREATE RESOURCE GROUP in TiDB. # SHOW CREATE RESOURCE GROUP -You can use the `SHOW CREATE RESOURCE GROUP` statement to view the current definition of a resource group, and recreate it in another TiDB cluster. +You can use the `SHOW CREATE RESOURCE GROUP` statement to view the current definition of a resource group. ## Synopsis diff --git a/tidb-resource-control.md b/tidb-resource-control.md index a2bb9db5bc156..3921699c75917 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -20,12 +20,28 @@ This feature applies to the following scenarios: In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. + + +> **Note:** +> +> This feature does not support [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). + + + ## Parameters for resource control The resource control feature introduces two new global variables. * TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control) to control whether to enable flow control for a resource group. -* TiKV: you can use the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. Note that this parameter does not support dynamic modification. You need to restart the TiKV instance to for the modification to take effect. +* TiKV: you can use the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. + + + +> **Note:** +> +> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. You need to restart the TiKV instance for the modification to take effect. + + The results of the combination of these two parameters are shown in the following table. @@ -36,3 +52,87 @@ The results of the combination of these two parameters are shown in the followin ## How to use resource control +For an existing resource group, you can modify the read and write quota of the resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md). The quota changes to the resource group take effect immediately. + +You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md). The user bound by the deleted resource group will use the `default` resource group for resource isolation. + +> **Note:** +> +> - When you bind a user resource group with `CREATE USER` or `ALTER USER`, it will not take effect for the user's existing sessions, but only for the user's new sessions. +> - The `default` resource group does not have quota restrictions for bound user applications. It is recommended to create the `default` resource group by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), or modify the quota for the `default` resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md) to control the quota for the `default` resource group. + +### Prerequisites + +To create, modify, or delete a resource group, you need to have the `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. + +You can create a resource group in the cluster with [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind the users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). + +### Step 1. Enable the resource control feature + +Enable flow control for the resource group. + +```sql +SET GLOBAL tidb_enable_resource_control = 'ON'; +``` + +In TiKV, set the parameter `resource_control.enabled` to `true`. The parameter `resource_control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. + +### Step 2. Create a resource group, and bind users to it + +Resource group quotas are expressed as [RU (Resource Unit)](/tidb-RU.md), which is TiDB's unified abstraction of CPU, IO, and other system resources. + +The following is an example of how to create a resource group and bind users to it. + +1. Create a resource group `rg1` with a quota of 500 RU per second for read requests and 300 RU per second for write requests, and allow applications in this resource group to use excessive resources when system resources are available. + + ```sql + CREATE RESOURCE GROUP IF NOT EXISTS rg1 + RRU_PER_SEC = 500 + WRU_PER_SEC = 300 + BURSTABLE + ; + ``` + +2. Create a resource group `rg2` with a quota of 600 RU per second for read requests and 400 RU per second for write requests, and do not allow applications in this resource group to use excessive resources even when system resources are available. + + ```sql + CREATE RESOURCE GROUP IF NOT EXISTS rg2 + RRU_PER_SEC = 600 + WRU_PER_SEC = 400 + ; + ``` + +3. Bind users `usr1` and `usr2` to resource groups `rg1` and `rg2` respectively. + + ```sql + ALTER USER usr1 RESOURCE GROUP rg1; + ``` + + ```sql + ALTER USER usr2 RESOURCE GROUP rg2; + ``` + +After you complete the preceeding operations, the resource consumption by newly created sessions is controlled by the specified quota. Read requests are limited by the quota of the read RU, and write requests are limited by the quota of the write RU. If the system load is relatively high and there is no spare capacity, the resource consumption rate of both users will be strictly controlled not to exceed the quota. Meanwhile, the consumption ratio of RU metrics for both users' read and write requests is basically proportional to the specified quota. When system resources are abundant, the resource consumption rate of `usr1` is allowed to exceed the quota because it has set `BURSTABLE`. + +## Monitoring and charts + +TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** panel. The metrics are detailed in the **Resource Control** section of [TiDB Important Monitoring Metrics](/grafana-tidb-dashboard.md). + +## Tool Compatibility + +The resource control feature is not compatible with data export/import and replication tools including BR, TiDB Lightning, and TiCDC. + +## Limitations + +Currently, the resource control feature has the following limitations: + +* This feature only supports restriction and scheduling of read and write requests initiated by frontend clients. It does not support restriction and scheduling of background tasks such as `DDL` and `Auto Analyze`. +* Resource control will incur additional scheduling overhead. Therefore, there might be a slight performance degradation when this feature is enabled. + +## See also + +* [CREATE RESOURCE GROUP](/sql-statements/sql-statement-create-resource-group.md) +* [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) +* [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) +* [RESOURCE GROUP RFC](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/) +* [RU](/tidb-RU.md) diff --git a/tikv-configuration-file.md b/tikv-configuration-file.md index 53cb61eaf1430..c5000985e37e7 100644 --- a/tikv-configuration-file.md +++ b/tikv-configuration-file.md @@ -2153,3 +2153,12 @@ To reduce write latency, TiKV periodically fetches and caches a batch of timesta + The maximum number of TSOs in a timestamp request. + In a default TSO physical time update interval (`50ms`), PD provides at most 262144 TSOs. When requested TSOs exceed this number, PD provides no more TSOs. This configuration item is used to avoid exhausting TSOs and the reverse impact of TSO exhaustion on other businesses. If you increase the value of this configuration item to improve high availability, you need to decrease the value of [`tso-update-physical-interval`](/pd-configuration-file.md#tso-update-physical-interval) at the same time to get enough TSOs. + Default value: `8192` + +## resource_control + +Configuration items related to resource control. + +### `enabled` New in v6.6.0 + ++ Control whether to use read/write request scheduling based on resource group quotas. ++ Default value: `false`, which means to disable scheduling based on resource group quotas. \ No newline at end of file From 59828dc211fc26e2348b9fabc4f241e59b5a0593 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 2 Feb 2023 11:32:07 +0800 Subject: [PATCH 08/63] removed blank line --- sql-statements/sql-statement-show-create-resource-group.md | 2 +- system-variables.md | 1 - 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/sql-statements/sql-statement-show-create-resource-group.md b/sql-statements/sql-statement-show-create-resource-group.md index a0e293a481468..031622acbbe0e 100644 --- a/sql-statements/sql-statement-show-create-resource-group.md +++ b/sql-statements/sql-statement-show-create-resource-group.md @@ -19,7 +19,7 @@ ResourceGroupName ::= ## Examples -View the current definition of the resource group `rg1` and recreate it in another TiDB cluster. +View the current definition of the resource group `rg1`. ```sql CREATE RESOURCE GROUP rg1 RRU_PER_SEC=100 WRU_PER_SEC=200; diff --git a/system-variables.md b/system-variables.md index 030ada0d437f7..98081189f355a 100644 --- a/system-variables.md +++ b/system-variables.md @@ -1836,7 +1836,6 @@ MPP is a distributed computing framework provided by the TiFlash engine, which a - Type: Boolean - This variable is a switch for [the resource control feature](/tidb-resource-control.md). When this variable is set to `ON`, the cluster supports resource isolation of applications by resource group. - ### tidb_enable_reuse_chunk New in v6.4.0 - Scope: SESSION | GLOBAL From c7f9e803794b9c598c596bfce82e13ae486663e5 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 2 Feb 2023 13:57:42 +0800 Subject: [PATCH 09/63] Apply suggestions from code review --- .../information-schema-resource-groups.md | 2 +- .../sql-statement-alter-resource-group.md | 6 ++--- .../sql-statement-create-resource-group.md | 2 +- tidb-resource-control.md | 23 +++++++++++-------- 4 files changed, 19 insertions(+), 14 deletions(-) diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index 9cc11c6a33a84..30de61f760f3f 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -3,7 +3,7 @@ title: RESOURCE_GROUPS summary: Learn the `RESOURCE_GROUPS` information_schema table. --- -The `RESOURCE_GROUPS` table shows information about all resource groups. See [Resource Conctrol in TiDB](/tidb-resource-control.md). +The `RESOURCE_GROUPS` table shows information about all resource groups. See [Use Resource Control to Achieve Resource Isolation](/tidb-resource-control.md). ```sql USE information_schema; diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 4ea13893c7c43..2fa605e889c2b 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -35,10 +35,10 @@ BurstableOption ::= TiDB supports the following `DirectResourceGroupOption`, where [RU (Resource Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. -| Parameter | Description | Example | +| Option | Description | Example | |---------------|-------------------------------------|------------------------| -|`RRU_PER_SEC` | Quota of RU read per second |`RRU_PER_SEC = 500` | -|`WRU_PER_SEC` | Quota of RU write per second |`WRU_PER_SEC = 300` | +|`RRU_PER_SEC` | Quota of read RU per second |`RRU_PER_SEC = 500` | +|`WRU_PER_SEC` | Quota of write RU per second |`WRU_PER_SEC = 300` | If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota if the system resources are sufficient. diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index e9bb21b3910ff..54c19fc09f0b1 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -65,7 +65,7 @@ mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; +------+--------------+---------------------------------------------------------------+ | Name | Plan_type | Directive | +------+--------------+---------------------------------------------------------------+ -| rg1 | tenancy | {"RRU_PER_SEC": 5000, "WRU_PER_SEC": 1000, "BURSTABLE": true} | +| rg1 | tenancy | {"RRU_PER_SEC": 500, "WRU_PER_SEC": 300, "BURSTABLE": true} | +------+--------------+---------------------------------------------------------------+ 1 row in set (0.00 sec) ``` diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 3921699c75917..4875e2a7cdccf 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -16,7 +16,7 @@ The introduction of the resource control feature is a milestone for TiDB. It can This feature applies to the following scenarios: - You can combine multiple small and medium-sized applications from different systems into one TiDB cluster. If the load of an individual application grows larger, it does not affect the normal operation of other businesses. When the system load is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. -- You can choose to combine all test environments into a single cluster, or group the batch tasks that consume more resources into a single resource group. It can improve hardware utilization and reduce operating costs while ensuring that critical applications receive the necessary resources. +- You can choose to combine all test environments into a single cluster, or group the batch tasks that consume more resources into a single resource group. It can improve hardware utilization and reduce operating costs while ensuring that critical applications can still get the necessary resources. In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. @@ -43,7 +43,7 @@ The resource control feature introduces two new global variables. -The results of the combination of these two parameters are shown in the following table. +The results of the combinations of these two parameters are shown in the following table. | `resource_control.enabled` | `tidb_enable_resource_control`= ON | `tidb_enable_resource_control`= OFF | |:----------------------------|:-------------------------------------|:-------------------------------------| @@ -54,18 +54,18 @@ The results of the combination of these two parameters are shown in the followin For an existing resource group, you can modify the read and write quota of the resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md). The quota changes to the resource group take effect immediately. -You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md). The user bound by the deleted resource group will use the `default` resource group for resource isolation. +You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md). The user bound to the deleted resource group will use the `default` resource group for resource isolation. > **Note:** > -> - When you bind a user resource group with `CREATE USER` or `ALTER USER`, it will not take effect for the user's existing sessions, but only for the user's new sessions. -> - The `default` resource group does not have quota restrictions for bound user applications. It is recommended to create the `default` resource group by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), or modify the quota for the `default` resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md) to control the quota for the `default` resource group. +> - When you bind a user to a resource group by using `CREATE USER` or `ALTER USER`, it will not take effect for the user's existing sessions, but only for the user's new sessions. +> - The `default` resource group does not have quota restrictions for the user's applications. It is recommended to create the `default` resource group by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), or modify the quota for the `default` resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md) to control the quota for the `default` resource group. ### Prerequisites To create, modify, or delete a resource group, you need to have the `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. -You can create a resource group in the cluster with [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind the users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). +You can create a resource group in the cluster by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind the users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). ### Step 1. Enable the resource control feature @@ -75,9 +75,14 @@ Enable flow control for the resource group. SET GLOBAL tidb_enable_resource_control = 'ON'; ``` + + + In TiKV, set the parameter `resource_control.enabled` to `true`. The parameter `resource_control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. -### Step 2. Create a resource group, and bind users to it + + +### Step 2. Create a resource group, and then bind users to it Resource group quotas are expressed as [RU (Resource Unit)](/tidb-RU.md), which is TiDB's unified abstraction of CPU, IO, and other system resources. @@ -112,7 +117,7 @@ The following is an example of how to create a resource group and bind users to ALTER USER usr2 RESOURCE GROUP rg2; ``` -After you complete the preceeding operations, the resource consumption by newly created sessions is controlled by the specified quota. Read requests are limited by the quota of the read RU, and write requests are limited by the quota of the write RU. If the system load is relatively high and there is no spare capacity, the resource consumption rate of both users will be strictly controlled not to exceed the quota. Meanwhile, the consumption ratio of RU metrics for both users' read and write requests is basically proportional to the specified quota. When system resources are abundant, the resource consumption rate of `usr1` is allowed to exceed the quota because it has set `BURSTABLE`. +After you complete the preceeding operations, the resource consumption by newly created sessions is controlled by the specified quota. Read requests are limited by the read RU quota, and write requests are limited by the write RU quota. If the system load is relatively high and there is no spare capacity, the resource consumption rate of both users will be strictly controlled not to exceed the quota. Meanwhile, the consumption ratio of RU metrics for both users' read and write requests is basically proportional to the specified quota. When system resources are abundant, the resource consumption rate of `usr1` is allowed to exceed the quota because it has set `BURSTABLE`, while `usr2` is not allowed. ## Monitoring and charts @@ -127,7 +132,7 @@ The resource control feature is not compatible with data export/import and repli Currently, the resource control feature has the following limitations: * This feature only supports restriction and scheduling of read and write requests initiated by frontend clients. It does not support restriction and scheduling of background tasks such as `DDL` and `Auto Analyze`. -* Resource control will incur additional scheduling overhead. Therefore, there might be a slight performance degradation when this feature is enabled. +* Resource control incurs additional scheduling overhead. Therefore, there might be a slight performance degradation when this feature is enabled. ## See also From 8218ab455a914363055ff9c609e3f62d0d2c672a Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 2 Feb 2023 14:04:22 +0800 Subject: [PATCH 10/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 4875e2a7cdccf..48c0c6b4ca64a 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -9,7 +9,7 @@ summary: Learn how to use resource management to control and schedule applicatio > > This feature is experimental and its form and usage may change in subsequent versions. -Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas by resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. +Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas by resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwhile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. From acc118354d7e3615b54bc033fe689372c37a16b3 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 2 Feb 2023 14:05:47 +0800 Subject: [PATCH 11/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 48c0c6b4ca64a..605e92d87557c 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -117,7 +117,7 @@ The following is an example of how to create a resource group and bind users to ALTER USER usr2 RESOURCE GROUP rg2; ``` -After you complete the preceeding operations, the resource consumption by newly created sessions is controlled by the specified quota. Read requests are limited by the read RU quota, and write requests are limited by the write RU quota. If the system load is relatively high and there is no spare capacity, the resource consumption rate of both users will be strictly controlled not to exceed the quota. Meanwhile, the consumption ratio of RU metrics for both users' read and write requests is basically proportional to the specified quota. When system resources are abundant, the resource consumption rate of `usr1` is allowed to exceed the quota because it has set `BURSTABLE`, while `usr2` is not allowed. +After you complete the preceding operations, the resource consumption by newly created sessions is controlled by the specified quota. Read requests are limited by the read RU quota, and write requests are limited by the write RU quota. If the system load is relatively high and there is no spare capacity, the resource consumption rate of both users will be strictly controlled not to exceed the quota. Meanwhile, the consumption ratio of RU metrics for both users' read and write requests is basically proportional to the specified quota. When system resources are abundant, the resource consumption rate of `usr1` is allowed to exceed the quota because it has set `BURSTABLE`, while `usr2` is not allowed. ## Monitoring and charts From dae0f0e93adae067505b2899cddf26f56573752f Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 2 Feb 2023 14:18:46 +0800 Subject: [PATCH 12/63] Apply suggestions from code review --- sql-statements/sql-statement-create-resource-group.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index 54c19fc09f0b1..ff3ee76fb4197 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -33,11 +33,11 @@ BurstableOption ::= ``` -The `ResourceGroupName` of a resource group is globally unique and cannot be duplicated. +The resource group name parameter (`ResourceGroupName`) is globally unique and cannot be duplicated. TiDB supports the following `DirectResourceGroupOption`, where [RU (Resource Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. -| Parameter | Description | Example | +| Option | Description | Example | |---------------|-------------------------------------|------------------------| |`RRU_PER_SEC` | Quota of RU read per second |`RRU_PER_SEC = 500` | |`WRU_PER_SEC` | Quota of RU write per second |`WRU_PER_SEC = 300` | From 79895e4555476d4e3d890b2f4b82a6f5f25ae8fe Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 3 Feb 2023 16:49:48 +0800 Subject: [PATCH 13/63] draft done --- grafana-resource-control-dashboard.md | 36 +++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) create mode 100644 grafana-resource-control-dashboard.md diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md new file mode 100644 index 0000000000000..ccc663f999f91 --- /dev/null +++ b/grafana-resource-control-dashboard.md @@ -0,0 +1,36 @@ +--- +title: Key Monitoring Metrics of Resource Control +summary: Learn some key metrics displayed on the Grafana Resource Control dashboard. +--- + +# Key Monitoring Metrics of Resource Control + +If you use TiUP to deploy the TiDB cluster, the monitoring system (Prometheus & Grafana) is deployed at the same time. For more information, see [Overview of the Monitoring Framework](/tidb-monitoring-framework.md). + +The Grafana dashboard is divided into a series of sub dashboards which include Overview, PD, TiDB, TiKV, Node\_exporter, Disk Performance, and Performance\_overview. A lot of metrics are there to help you diagnose. + +You can get an overview of the resource control status from the PD dashboard, where the key metrics are displayed. This document provides a detailed description of these key metrics. + +The following is the description of Resource Control Dashboard metrics items. + +## Resource Unit + +- RU: the [Resource Unit](/tidb-RU.md) consumption information of each Resource Group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each Resource Group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). +- RU Per Query: the average number of Request Units consumed by each SQL statement per second is obtained by dividing the above Request Unit by the number of SQL statements executed per second. +- RRU: the Read Request Unit consumption information of each Resource Group is calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. +- RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second is obtained by dividing the above Read Request Unit by the number of SQL statements executed per second. +- WRU: the Write Request Unit consumption information of each Resource Group is calculated in real time. `total` is the sum of the Write Request Units consumed by all Resource Groups. +- WRU Per Query: the average number of Write Request Units consumed by each SQL statement per second is obtained by dividing the above Write Request Unit by the number of SQL statements executed per second. + +## Resource + +- KV Request Count: the number of KV requests is calculated in real time for each Resource Group. They are divided into read and write types. `total` is the sum of the KV requests involved by all Resource Groups. +- KV Request Count Per Query: the average number of read and write KV requests by each SQL statement per second is obtained by dividing the above KV Request Count by the number of SQL statements executed per second. +- Bytes Read: the amount of data read is calculated in real time for each Resource Group. `total` is the sum of the data read by all Resource Groups. +- Bytes Read Per Query: the average amount of data read by each SQL statement per second is obtained by dividing the above Bytes Read by the number of SQL statements executed per second. +- Bytes Written: the amount of data written is calculated in real time for each Resource Group. `total` is the sum of the data written by all Resource Groups. +- Bytes Written Per Query: the average amount of data written by each SQL statement per second is obtained by dividing the above Bytes Written by the number of SQL statements executed per second. +- KV CPU Time: the KV layer CPU time consumed is calculated in real time for each Resource Group. `total` is the sum of the KV layer CPU time consumed by all Resource Groups. +- KV CPU Time Per Query: the average KV layer CPU time consumed by each SQL statement per second is obtained by dividing the above KV CPU Time by the number of SQL statements executed per second. +- SQL CPU Time: the SQL layer CPU time consumed is calculated in real time for each Resource Group. `total` is the sum of the SQL layer CPU time consumed by all Resource Groups. +- SQL CPU Time Per Query: the average SQL layer CPU time consumed by each SQL statement per second is obtained by dividing the above SQL CPU Time by the number of SQL statements executed per second. From 32cd4c241d597c86508be153114300b1045fc47a Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 3 Feb 2023 16:55:30 +0800 Subject: [PATCH 14/63] Update tidb-resource-control.md --- tidb-resource-control.md | 1 - 1 file changed, 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 605e92d87557c..87efa1311d84c 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -75,7 +75,6 @@ Enable flow control for the resource group. SET GLOBAL tidb_enable_resource_control = 'ON'; ``` - In TiKV, set the parameter `resource_control.enabled` to `true`. The parameter `resource_control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. From f5a5a9f3dfe643ab1ad6262264c2ced4a53d24f1 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 3 Feb 2023 21:51:42 +0800 Subject: [PATCH 15/63] sync brian's latest updates --- .../information-schema-processlist.md | 35 +++++----- .../information-schema-resource-groups.md | 68 ++++++------------- .../sql-statement-alter-resource-group.md | 46 ++++++------- .../sql-statement-create-resource-group.md | 37 +++++----- .../sql-statement-drop-resource-group.md | 37 ++++------ ...ql-statement-show-create-resource-group.md | 4 +- tidb-resource-control.md | 57 +++++++++------- tikv-configuration-file.md | 5 +- 8 files changed, 131 insertions(+), 158 deletions(-) diff --git a/information-schema/information-schema-processlist.md b/information-schema/information-schema-processlist.md index 934448b124026..510d056bdebf8 100644 --- a/information-schema/information-schema-processlist.md +++ b/information-schema/information-schema-processlist.md @@ -13,7 +13,7 @@ The `PROCESSLIST` table has additional columns not present in `SHOW PROCESSLIST` * A `MEM` column to show the memory used by the request that is being processed, in bytes. * A `DISK` column to show the disk usage in bytes. * A `TxnStart` column to show the start time of the transaction. -* A `RESOURCE_GROUP_NAME` column to show the resource group name. +* A `RESOURCE_GROUP` column to show the resource group name. {{< copyable "sql" >}} @@ -38,7 +38,7 @@ DESC processlist; | MEM | bigint(21) unsigned | YES | | NULL | | | DISK | bigint(21) unsigned | YES | | NULL | | | TxnStart | varchar(64) | NO | | | | -| RESOURCE_GROUP_NAME | varchar(32) | NO | | | | +| RESOURCE_GROUP | varchar(32) | NO | | | | +---------------------+---------------------+------+------+---------+-------+ 13 rows in set (0.00 sec) ``` @@ -48,13 +48,12 @@ DESC processlist; ```sql SELECT * FROM processlist\G ``` - ```sql *************************** 1. row *************************** - ID: 5794501045170209171 + ID: 2300033189772525975 USER: root - HOST: 127.0.0.1:60645 - DB: information_schema + HOST: 127.0.0.1:51289 + DB: NULL COMMAND: Query TIME: 0 STATE: autocommit @@ -63,7 +62,7 @@ SELECT * FROM processlist\G MEM: 0 DISK: 0 TxnStart: -RESOURCE_GROUP_NAME: default + RESOURCE_GROUP: rg1 1 row in set (0.00 sec) ``` @@ -81,7 +80,7 @@ Fields in the `PROCESSLIST` table are described as follows: * MEM: The memory used by the request that is being processed, in bytes. * DISK: The disk usage in bytes. * TxnStart: The start time of the transaction. -* RESOURCE_GROUP_NAME: The resource group name. +* RESOURCE_GROUP: The resource group name. ## CLUSTER_PROCESSLIST @@ -94,14 +93,14 @@ SELECT * FROM information_schema.cluster_processlist; ``` ```sql -+-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+---------------------+ -| INSTANCE | ID | USER | HOST | DB | COMMAND | TIME | STATE | INFO | MEM | TxnStart | RESOURCE_GROUP_NAME | -+-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+---------------------+ - -| 10.0.1.22:10080 | 150 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077223) | default | -| 10.0.1.22:10080 | 138 | root | 10.0.1.1 | test | Query | 0 | autocommit | SELECT * FROM information_schema.cluster_processlist | 0 | 05-28 03:54:21.230(416976223923077220) | rg1 | -| 10.0.1.22:10080 | 151 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077224) | rg2 | -| 10.0.1.21:10080 | 15 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077222) | default | -| 10.0.1.21:10080 | 14 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077225) | default | -+-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+---------------------+ ++-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+----------------+ +| INSTANCE | ID | USER | HOST | DB | COMMAND | TIME | STATE | INFO | MEM | TxnStart | RESOURCE_GROUP | ++-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+----------------+ + +| 10.0.1.22:10080 | 150 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077223) | default | +| 10.0.1.22:10080 | 138 | root | 10.0.1.1 | test | Query | 0 | autocommit | SELECT * FROM information_schema.cluster_processlist | 0 | 05-28 03:54:21.230(416976223923077220) | rg1 | +| 10.0.1.22:10080 | 151 | u1 | 10.0.1.1 | test | Query | 0 | autocommit | select count(*) from usertable | 372 | 05-28 03:54:21.230(416976223923077224) | rg2 | +| 10.0.1.21:10080 | 15 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077222) | default | +| 10.0.1.21:10080 | 14 | u2 | 10.0.1.1 | test | Query | 0 | autocommit | select max(field0) from usertable | 496 | 05-28 03:54:21.230(416976223923077225) | default | ++-----------------+-----+------+----------+------+---------+------+------------+------------------------------------------------------+-----+----------------------------------------+----------------+ ``` diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index 30de61f760f3f..cfd15e501e332 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -11,20 +11,15 @@ DESC resource_groups; ``` ```sql -+-----------------+-------------+------+------+---------+-------+ -| Field | Type | Null | Key | Default | Extra | -+-----------------+-------------+------+------+---------+-------+ -| NAME | varchar(32) | NO | | NULL | | -| MODE | varchar(12) | NO | | NULL | | -| RRU | bigint(21) | YES | | NULL | | -| RRU_TOKEN | bigint(21) | YES | | NULL | | -| WRU | bigint(21) | YES | | NULL | | -| WRU_TOKEN | bigint(21) | YES | | NULL | | -| CPU | bigint(21) | YES | | NULL | | -| READ_BANDWIDTH | bigint(21) | YES | | NULL | | -| WRITE_BANDWIDTH | bigint(21) | YES | | NULL | | -+-----------------+-------------+------+------+---------+-------+ -9 rows in set (0.08 sec) ++------------+-------------+------+------+---------+-------+ +| Field | Type | Null | Key | Default | Extra | ++------------+-------------+------+------+---------+-------+ +| NAME | varchar(32) | NO | | NULL | | +| RU_PER_SEC | bigint(21) | YES | | NULL | | +| RU_TOKENS | bigint(21) | YES | | NULL | | +| BURSTABLE | varchar(3) | YES | | NULL | | ++------------+-------------+------+------+---------+-------+ +4 rows in set (0.00 sec) ``` ## Examples @@ -32,39 +27,18 @@ DESC resource_groups; {{< copyable "sql" >}} ```sql -mysql> CREATE RESOURCE GROUP rg1 RRU_PER_SEC=1000 WRU_PER_SEC=2000 ; -- Use RU_MODE to create the resource group rg1 +mysql> CREATE RESOURCE GROUP rg1 RU_PER_SEC=1000; -- Create the resource group rg1 Query OK, 0 rows affected (0.34 sec) mysql> SHOW CREATE RESOURCE GROUP rg1; -- Displays the definition of the rg1 resource group. -+----------------+------------------------------------------------------------------+ -| Resource_Group | Create Resource Group | -+----------------+------------------------------------------------------------------+ -| rg1 | CREATE RESOURCE GROUP `rg1` RRU_PER_SEC=1000 WRU_PER_SEC=2000 | -+----------------+------------------------------------------------------------------+ ++----------------+---------------------------------------------+ +| Resource_Group | Create Resource Group | ++----------------+---------------------------------------------+ +| rg1 | CREATE RESOURCE GROUP `rg1` RU_PER_SEC=1000 | ++----------------+---------------------------------------------+ 1 row in set (0.00 sec) -mysql> SELECT * FROM information_schema.placement_policies WHERE NAME = 'rg1'; -- Displays the runtime status of the resource group rg1. -+--------+---------+------+-----------+------+-----------+------+----------------+-----------------+ -| NAME | MODE | RRU | RRU_TOKEN | WRU | WRU_TOKEN | CPU | READ_BANDWIDTH | WRITE_BANDWIDTH | -+--------+---------+------+-----------+------+-----------+------+----------------+-----------------+ -| rg1 | RU_MODE | 1000 | 168999 | 2000 | 249999 | NULL | NULL | NULL | -+--------+---------+------+-----------+------+-----------+------+----------------+-----------------+ -1 row in set (0.02 sec) -``` - -```sql -mysql> CREATE RESOURCE GROUP rg2 CPU="8" IO_READ_BANDWIDTH="1000Mi" IO_WRITE_BANDWIDTH="1000Mi"; -- Use RAW_MODE to create the resource group rg2 -Query OK, 0 rows affected (0.20 sec) -mysql> show create resource group rg2; -- Displays the definition of the rg2 resource group. -+----------------+-----------------------------------------------------------------------------------------------+ -| Resource_Group | Create Resource Group | -+----------------+-----------------------------------------------------------------------------------------------+ -| rg2 | CREATE RESOURCE GROUP `rg2` CPU="8" IO_READ_BANDWIDTH="1000Mi" IO_WRITE_BANDWIDTH="1000Mi" | -+----------------+-----------------------------------------------------------------------------------------------+ -1 row in set (0.00 sec) -mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg2'; -- Displays the runtime status of the resource group rg2. -+--------+----------+------+-----------+------+-----------+------+----------------+-----------------+ -| NAME | MODE | RRU | RRU_TOKEN | WRU | WRU_TOKEN | CPU | READ_BANDWIDTH | WRITE_BANDWIDTH | -+--------+----------+------+-----------+------+-----------+------+----------------+-----------------+ -| rg2 | RAW_MODE | NULL | NULL | NULL | NULL | 8000 | 1048576000 | 1048576000 | -+--------+----------+------+-----------+------+-----------+------+----------------+-----------------+ -1 row in set (0.00 sec) -``` +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg1'; ++------+------------+-----------+-----------+ +| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | ++------+------------+-----------+-----------+ +| rg1 | 1000 | 100000 | NO | ++------+------------+-----------+-----------+ diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 2fa605e889c2b..29da65a9d9961 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -33,14 +33,13 @@ BurstableOption ::= ``` -TiDB supports the following `DirectResourceGroupOption`, where [RU (Resource Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. +TiDB supports the following `DirectResourceGroupOption`, where [RU (Request Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. | Option | Description | Example | |---------------|-------------------------------------|------------------------| -|`RRU_PER_SEC` | Quota of read RU per second |`RRU_PER_SEC = 500` | -|`WRU_PER_SEC` | Quota of write RU per second |`WRU_PER_SEC = 300` | +|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RU per second | -If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota if the system resources are sufficient. +If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota. > **Note:** > @@ -54,31 +53,26 @@ Create a resource group named `rg1` and modify its properties. mysql> DROP RESOURCE GROUP IF EXISTS rg1; Query OK, 0 rows affected (0.22 sec) mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 - -> RRU_PER_SEC = 500 - -> WRU_PER_SEC = 300 - -> BURSTABLE - -> ; + -> RU_PER_SEC = 100 + -> BURSTABLE; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+--------------+---------------------------------------------------------------+ -| Name | Plan_type | Directive | -+------+--------------+---------------------------------------------------------------+ -| rg1 | tenancy | {"RRU_PER_SEC": 500, "WRU_PER_SEC": 300, "BURSTABLE": true} | -+------+--------------+---------------------------------------------------------------+ -1 row in set (0.00 sec) - -mysql> ALTER RESOURCE GROUP IF NOT EXISTS rg1 - -> RRU_PER_SEC = 600 - -> WRU_PER_SEC = 400 - -> ; -Query OK, 0 rows affected (0.09 sec) ++------+-------------+-----------+-----------+ +| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | ++------+-------------+-----------+-----------+ +| rg1 | 100 | 165135 | YES | ++------+-------------+-----------+-----------+ +1 rows in set (1.30 sec) +mysql> ALTER RESOURCE GROUP rg1 + -> RU_PER_SEC = 200; +Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+--------------+---------------------------------------------------------------+ -| Name | Plan_type | Directive | -+------+--------------+---------------------------------------------------------------+ -| rg1 | tenancy | {"RRU_PER_SEC": 600, "WRU_PER_SEC": 400, "BURSTABLE": false} | -+------+--------------+---------------------------------------------------------------+ -1 row in set (0.00 sec) ++------+-------------+-----------+-----------+ +| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | ++------+-------------+-----------+-----------+ +| rg1 | 200 | 257158 | NO | ++------+-------------+-----------+-----------+ +1 rows in set (1.30 sec) ``` ## MySQL compatibility diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index ff3ee76fb4197..a6bf053e2cca0 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -25,8 +25,7 @@ ResourceGroupOptionList: | ResourceGroupOptionList ',' DirectResourceGroupOption DirectResourceGroupOption: - "RRU_PER_SEC" EqOpt stringLit -| "WRU_PER_SEC" EqOpt stringLit + "RU_PER_SEC" EqOpt stringLit BurstableOption ::= ("BURSTABLE")? @@ -35,14 +34,13 @@ BurstableOption ::= The resource group name parameter (`ResourceGroupName`) is globally unique and cannot be duplicated. -TiDB supports the following `DirectResourceGroupOption`, where [RU (Resource Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. +TiDB supports the following `DirectResourceGroupOption`, where [RU (Request Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. | Option | Description | Example | |---------------|-------------------------------------|------------------------| -|`RRU_PER_SEC` | Quota of RU read per second |`RRU_PER_SEC = 500` | -|`WRU_PER_SEC` | Quota of RU write per second |`WRU_PER_SEC = 300` | +|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RU per second | -If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota if the system resources are sufficient. +If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota. > **Note:** > @@ -50,24 +48,27 @@ If the `BURSTABLE` attribute is set, the corresponding resource group allows the ## Examples -Create a resource group named `rg1`. +Create two resource groups `rg1` and `rg2`. ```sql mysql> DROP RESOURCE GROUP IF EXISTS rg1; Query OK, 0 rows affected (0.22 sec) mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 - -> RRU_PER_SEC = 500 - -> WRU_PER_SEC = 300 - -> BURSTABLE - -> ; + -> RU_PER_SEC = 100 + -> BURSTABLE; Query OK, 0 rows affected (0.08 sec) -mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+--------------+---------------------------------------------------------------+ -| Name | Plan_type | Directive | -+------+--------------+---------------------------------------------------------------+ -| rg1 | tenancy | {"RRU_PER_SEC": 500, "WRU_PER_SEC": 300, "BURSTABLE": true} | -+------+--------------+---------------------------------------------------------------+ -1 row in set (0.00 sec) +mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg2 + -> RU_PER_SEC = 200 + -> BURSTABLE; +Query OK, 0 rows affected (0.08 sec) +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1' or NAME = 'rg2'; ++------+-------------+-----------+-----------+ +| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | ++------+-------------+-----------+-----------+ +| rg1 | 100 | 165135 | YES | +| rg2 | 200 | 157158 | NO | ++------+-------------+-----------+-----------+ +2 rows in set (1.30 sec) ``` ## MySQL compatibility diff --git a/sql-statements/sql-statement-drop-resource-group.md b/sql-statements/sql-statement-drop-resource-group.md index 9c68334564084..47d1e01bfe8bc 100644 --- a/sql-statements/sql-statement-drop-resource-group.md +++ b/sql-statements/sql-statement-drop-resource-group.md @@ -11,10 +11,10 @@ You can use the `DROP RESOURCE GROUP` statement to drop a resource group in the ```ebnf+diagram DropResourceGroupStmt: - "DROP" "RESOURCE" "GROUP" IfNotExists ResourceGroupName + "DROP" "RESOURCE" "GROUP" IfExists ResourceGroupName -IfNotExists ::= - ('IF' 'NOT' 'EXISTS')? +IfExists ::= + ('IF' 'EXISTS')? ResourceGroupName: Identifier @@ -31,28 +31,21 @@ Drop a resource group named `rg1`. ```sql mysql> DROP RESOURCE GROUP IF EXISTS rg1; Query OK, 0 rows affected (0.22 sec) -mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 ( - -> RRU_PER_SEC = 500 - -> WRU_PER_SEC = 300 - -> BURSTABLE - -> ); +mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 RU_PER_SEC = 500 BURSTABLE; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+--------------+---------------------------------------------------------------+ -| Name | Plan_type | Directive | -+------+--------------+---------------------------------------------------------------+ -| rg1 | tenancy | {"RRU_PER_SEC": 500, "WRU_PER_SEC": 300, "BURSTABLE": true} | -+------+--------------+---------------------------------------------------------------+ -1 row in set (0.00 sec) - -mysql> DROP RESOURCE GROUP IF NOT EXISTS rg1 ; ++------+------------+-----------+-----------+ +| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | ++------+------------+-----------+-----------+ +| rg1 | 500 | 0 | YES | ++------+------------+-----------+-----------+ +1 row in set (0.01 sec) + +mysql> DROP RESOURCE GROUP IF EXISTS rg1 ; Query OK, 1 rows affected (0.09 sec) -mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+--------------+---------------------------------------------------------------+ -| Name | Plan_type | Directive | -+------+--------------+---------------------------------------------------------------+ -+------+--------------+---------------------------------------------------------------+ -0 row in set (0.00 sec) + +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; +Empty set (0.00 sec) ``` ## MySQL compatibility diff --git a/sql-statements/sql-statement-show-create-resource-group.md b/sql-statements/sql-statement-show-create-resource-group.md index 031622acbbe0e..e3a8ae72071c3 100644 --- a/sql-statements/sql-statement-show-create-resource-group.md +++ b/sql-statements/sql-statement-show-create-resource-group.md @@ -22,7 +22,7 @@ ResourceGroupName ::= View the current definition of the resource group `rg1`. ```sql -CREATE RESOURCE GROUP rg1 RRU_PER_SEC=100 WRU_PER_SEC=200; +CREATE RESOURCE GROUP rg1 RU_PER_SEC=100; Query OK, 0 rows affected (0.10 sec) ``` @@ -30,7 +30,7 @@ Query OK, 0 rows affected (0.10 sec) SHOW CREATE RESOURCE GROUP rg1; ***************************[ 1. row ]*************************** Resource_Group | rg1 -Create Resource Group | CREATE RESOURCE GROUP `rg1` RRU_PER_SEC=100 WRU_PER_SEC=200 +Create Resource Group | CREATE RESOURCE GROUP `rg1` RU_PER_SEC=100 1 row in set (0.00 sec) ``` diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 87efa1311d84c..14c2b717fce60 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -28,12 +28,30 @@ In addition, the rational use of the resource control feature can reduce the num +## What is Request Unit (RU) + +Request Unit (RU) is TiDB's unified abstraction unit for CPU, IO and other system resources, which currently includes CPU, IOPS and IO bandwidth metrics. The consumption of these three metrics is represented by the RU unit according to a certain ratio. + +The following table shows the consumption of TiKV storage layer CPU and IO resources by user requests and the corresponding RU weights. + +| Resource | RU Weight | +|:----------|:------| +| CPU | 1 RU / Milliseconds | +| Read IO | 1 RU / MiB | +| Write IO | 5 RU / MiB | +| Basic overhead of a read request | 1 RU | +| Basic overhead of a write request | 3 RU | + +Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` MiB data, and `w1` write requests that write `w2` MiB data, then the formula for the total RUs consumed by the resource group is as follows: + +c + (r1 + r2) + (3 * w1 + 5 * w2) + ## Parameters for resource control The resource control feature introduces two new global variables. -* TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control) to control whether to enable flow control for a resource group. -* TiKV: you can use the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. +* TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control) to control whether to enable flow control for a resource group. +* TiKV: you can use the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. @@ -52,20 +70,18 @@ The results of the combinations of these two parameters are shown in the followi ## How to use resource control +To create, modify, or delete a resource group, you need to have the `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. + +You can create a resource group in the cluster by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind the users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). + For an existing resource group, you can modify the read and write quota of the resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md). The quota changes to the resource group take effect immediately. -You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md). The user bound to the deleted resource group will use the `default` resource group for resource isolation. +You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md). > **Note:** > > - When you bind a user to a resource group by using `CREATE USER` or `ALTER USER`, it will not take effect for the user's existing sessions, but only for the user's new sessions. -> - The `default` resource group does not have quota restrictions for the user's applications. It is recommended to create the `default` resource group by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), or modify the quota for the `default` resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md) to control the quota for the `default` resource group. - -### Prerequisites - -To create, modify, or delete a resource group, you need to have the `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. - -You can create a resource group in the cluster by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind the users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). +> - If a user is not bound to a resource group or is bound to a `default` resource group, the user's requests are not subject to TiDB's flow control restrictions. The `default` resource group is currently not visible to the user and cannot be created or modified. ### Step 1. Enable the resource control feature @@ -83,27 +99,20 @@ In TiKV, set the parameter `resource_control.enabled` to `true`. The parameter ` ### Step 2. Create a resource group, and then bind users to it -Resource group quotas are expressed as [RU (Resource Unit)](/tidb-RU.md), which is TiDB's unified abstraction of CPU, IO, and other system resources. +Resource group quotas are expressed as [RU (Request Unit)](/tidb-RU.md), which is TiDB's unified abstraction of CPU, IO, and other system resources. The following is an example of how to create a resource group and bind users to it. -1. Create a resource group `rg1` with a quota of 500 RU per second for read requests and 300 RU per second for write requests, and allow applications in this resource group to use excessive resources when system resources are available. +1. Create a resource group `rg1`. The RU backfill rate is 500 RU per second and allows applications in this resource group to overrun resources. ```sql - CREATE RESOURCE GROUP IF NOT EXISTS rg1 - RRU_PER_SEC = 500 - WRU_PER_SEC = 300 - BURSTABLE - ; + CREATE RESOURCE GROUP IF NOT EXISTS rg1 RU_PER_SEC = 500 BURSTABLE; ``` -2. Create a resource group `rg2` with a quota of 600 RU per second for read requests and 400 RU per second for write requests, and do not allow applications in this resource group to use excessive resources even when system resources are available. +2. Create a resource group `rg2`. The RU backfill rate is 500 RU per second and does not allow applications in this resource group to overrun resources. ```sql - CREATE RESOURCE GROUP IF NOT EXISTS rg2 - RRU_PER_SEC = 600 - WRU_PER_SEC = 400 - ; + CREATE RESOURCE GROUP IF NOT EXISTS rg2 RU_PER_SEC = 600; ``` 3. Bind users `usr1` and `usr2` to resource groups `rg1` and `rg2` respectively. @@ -116,7 +125,9 @@ The following is an example of how to create a resource group and bind users to ALTER USER usr2 RESOURCE GROUP rg2; ``` -After you complete the preceding operations, the resource consumption by newly created sessions is controlled by the specified quota. Read requests are limited by the read RU quota, and write requests are limited by the write RU quota. If the system load is relatively high and there is no spare capacity, the resource consumption rate of both users will be strictly controlled not to exceed the quota. Meanwhile, the consumption ratio of RU metrics for both users' read and write requests is basically proportional to the specified quota. When system resources are abundant, the resource consumption rate of `usr1` is allowed to exceed the quota because it has set `BURSTABLE`, while `usr2` is not allowed. +After you complete the above operations of creating resource groups and binding users, the resource consumption of newly created sessions will be controlled by the specified quota. If the system load is relatively high and there is no spare capacity, the resource consumption rate of `usr2` will be strictly controlled not to exceed the quota. Because `usr1` bound by `rg1` is configured with `BURSTABLE`, the consumption rate of `usr1` is allowed to exceed the quota. + +If the resource group corresponds to a request with insufficient quota, the client's request will wait. If the wait time is too long, the request will report an error. ## Monitoring and charts diff --git a/tikv-configuration-file.md b/tikv-configuration-file.md index c5000985e37e7..4e110b44a03c9 100644 --- a/tikv-configuration-file.md +++ b/tikv-configuration-file.md @@ -2156,9 +2156,10 @@ To reduce write latency, TiKV periodically fetches and caches a batch of timesta ## resource_control -Configuration items related to resource control. +Configuration items related to resource control related to the TiKV storage layer. ### `enabled` New in v6.6.0 -+ Control whether to use read/write request scheduling based on resource group quotas. ++ Whether to support prioritized scheduling for user foreground read/write requests according to the corresponding resource group quota. For information about TiDB resource groups and resource control, see [TiDB resource control](/tidb-resource-control.md) ++ It is only meaningful to enable this configuration item if [`tidb_enable_resource_control](/system-variables.md#tidb_enable_resource_control-new-in-v660) is enabled on TiDB side. When it is enabled, TiKV will use the priority queue to schedule the queued requests for user foreground read/write requests. The priority of scheduling is inversely related to the amount of resources already consumed by the resource group where the request is located, and positively related to the quota of the corresponding resource group. + Default value: `false`, which means to disable scheduling based on resource group quotas. \ No newline at end of file From f85c858b1cd23a3acbd178d8bcf6054a15d6d28c Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Mon, 6 Feb 2023 09:31:57 +0800 Subject: [PATCH 16/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 14c2b717fce60..5c5b69fe363cc 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -50,7 +50,7 @@ c + (r1 + r2) + (3 * w1 + 5 * w2) The resource control feature introduces two new global variables. -* TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control) to control whether to enable flow control for a resource group. +* TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) to control whether to enable flow control for a resource group. * TiKV: you can use the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. From ed9f08d360f70e8e8472998ef976a976245f97d4 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Mon, 6 Feb 2023 09:35:20 +0800 Subject: [PATCH 17/63] Apply suggestions from code review Co-authored-by: BornChanger <97348524+BornChanger@users.noreply.github.com> --- grafana-resource-control-dashboard.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index ccc663f999f91..c444c96d62462 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -13,9 +13,9 @@ You can get an overview of the resource control status from the PD dashboard, wh The following is the description of Resource Control Dashboard metrics items. -## Resource Unit +## Request Unit -- RU: the [Resource Unit](/tidb-RU.md) consumption information of each Resource Group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each Resource Group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). +- RU: the [Request Unit](/tidb-RU.md) consumption information of each Resource Group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each Resource Group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). - RU Per Query: the average number of Request Units consumed by each SQL statement per second is obtained by dividing the above Request Unit by the number of SQL statements executed per second. - RRU: the Read Request Unit consumption information of each Resource Group is calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. - RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second is obtained by dividing the above Read Request Unit by the number of SQL statements executed per second. From 4571a11573e9d6950f6ebfa28243807c356b94b3 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Mon, 6 Feb 2023 10:22:26 +0800 Subject: [PATCH 18/63] fix links --- grafana-resource-control-dashboard.md | 2 +- sql-statements/sql-statement-alter-resource-group.md | 4 ++-- sql-statements/sql-statement-create-resource-group.md | 4 ++-- sql-statements/sql-statement-drop-resource-group.md | 2 +- tidb-resource-control.md | 4 ++-- 5 files changed, 8 insertions(+), 8 deletions(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index c444c96d62462..16dd982daaf43 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -15,7 +15,7 @@ The following is the description of Resource Control Dashboard metrics items. ## Request Unit -- RU: the [Request Unit](/tidb-RU.md) consumption information of each Resource Group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each Resource Group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). +- RU: the [Request Unit](/tidb-resource-control.md#what-is-request-unit-ru) consumption information of each Resource Group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each Resource Group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). - RU Per Query: the average number of Request Units consumed by each SQL statement per second is obtained by dividing the above Request Unit by the number of SQL statements executed per second. - RRU: the Read Request Unit consumption information of each Resource Group is calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. - RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second is obtained by dividing the above Read Request Unit by the number of SQL statements executed per second. diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 29da65a9d9961..37d84d0edefbc 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -33,7 +33,7 @@ BurstableOption ::= ``` -TiDB supports the following `DirectResourceGroupOption`, where [RU (Request Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. +TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) is TiDB's unified abstraction unit for CPU, IO, and other system resources. | Option | Description | Example | |---------------|-------------------------------------|------------------------| @@ -83,4 +83,4 @@ MySQL also supports [ALTER RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/ * [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) * [CREATE RESOURCE GROUP](/sql-statements/sql-statement-create-resource-group.md) -* [RU](/tidb-RU.md) +* [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index a6bf053e2cca0..eb67b8fa45ea5 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -34,7 +34,7 @@ BurstableOption ::= The resource group name parameter (`ResourceGroupName`) is globally unique and cannot be duplicated. -TiDB supports the following `DirectResourceGroupOption`, where [RU (Request Unit)](/tidb-RU.md) is TiDB's unified abstraction unit for CPU, IO, and other system resources. +TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) is TiDB's unified abstraction unit for CPU, IO, and other system resources. | Option | Description | Example | |---------------|-------------------------------------|------------------------| @@ -79,4 +79,4 @@ MySQL also supports [CREATE RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0 * [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) * [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) -* [RU](/tidb-RU.md) +* [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) diff --git a/sql-statements/sql-statement-drop-resource-group.md b/sql-statements/sql-statement-drop-resource-group.md index 47d1e01bfe8bc..ae22d67263394 100644 --- a/sql-statements/sql-statement-drop-resource-group.md +++ b/sql-statements/sql-statement-drop-resource-group.md @@ -56,4 +56,4 @@ MySQL also supports [DROP RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/e * [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) * [CREATE RESOURCE GROUP](/sql-statements/sql-statement-create-resource-group.md) -* [RU](/tidb-RU.md) \ No newline at end of file +* [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) \ No newline at end of file diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 5c5b69fe363cc..810e4de67ff77 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -99,7 +99,7 @@ In TiKV, set the parameter `resource_control.enabled` to `true`. The parameter ` ### Step 2. Create a resource group, and then bind users to it -Resource group quotas are expressed as [RU (Request Unit)](/tidb-RU.md), which is TiDB's unified abstraction of CPU, IO, and other system resources. +Resource group quotas are expressed as [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru), which is TiDB's unified abstraction of CPU, IO, and other system resources. The following is an example of how to create a resource group and bind users to it. @@ -150,4 +150,4 @@ Currently, the resource control feature has the following limitations: * [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) * [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) * [RESOURCE GROUP RFC](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/) -* [RU](/tidb-RU.md) +* [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) From 8489abc664e6c02d44909dd0405a61a67f65d3a9 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Mon, 6 Feb 2023 10:40:12 +0800 Subject: [PATCH 19/63] Update tidb-resource-control.md --- tidb-resource-control.md | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 810e4de67ff77..ea3968409e510 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -79,7 +79,7 @@ For an existing resource group, you can modify the read and write quota of the r You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md). > **Note:** -> +> > - When you bind a user to a resource group by using `CREATE USER` or `ALTER USER`, it will not take effect for the user's existing sessions, but only for the user's new sessions. > - If a user is not bound to a resource group or is bound to a `default` resource group, the user's requests are not subject to TiDB's flow control restrictions. The `default` resource group is currently not visible to the user and cannot be created or modified. @@ -99,8 +99,6 @@ In TiKV, set the parameter `resource_control.enabled` to `true`. The parameter ` ### Step 2. Create a resource group, and then bind users to it -Resource group quotas are expressed as [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru), which is TiDB's unified abstraction of CPU, IO, and other system resources. - The following is an example of how to create a resource group and bind users to it. 1. Create a resource group `rg1`. The RU backfill rate is 500 RU per second and allows applications in this resource group to overrun resources. @@ -150,4 +148,3 @@ Currently, the resource control feature has the following limitations: * [ALTER RESOURCE GROUP](/sql-statements/sql-statement-alter-resource-group.md) * [DROP RESOURCE GROUP](/sql-statements/sql-statement-drop-resource-group.md) * [RESOURCE GROUP RFC](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/) -* [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) From e7f6fab7930d8b6e537530e351fcd90f4502a97a Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Mon, 6 Feb 2023 22:43:03 +0800 Subject: [PATCH 20/63] translated the newly added content --- .../information-schema-resource-groups.md | 13 +++++++++++++ .../sql-statement-alter-resource-group.md | 13 +++++-------- .../sql-statement-create-resource-group.md | 7 ++----- tidb-resource-control.md | 9 ++++++++- 4 files changed, 28 insertions(+), 14 deletions(-) diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index cfd15e501e332..165a09ccae692 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -3,6 +3,12 @@ title: RESOURCE_GROUPS summary: Learn the `RESOURCE_GROUPS` information_schema table. --- +# RESOURCE_GROUPS + +> **Warning:** +> +> This feature is experimental and its form and usage may change in subsequent versions. + The `RESOURCE_GROUPS` table shows information about all resource groups. See [Use Resource Control to Achieve Resource Isolation](/tidb-resource-control.md). ```sql @@ -42,3 +48,10 @@ mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg1'; +------+------------+-----------+-----------+ | rg1 | 1000 | 100000 | NO | +------+------------+-----------+-----------+ + +The descriptions of the columns in the `RESOURCE_GROUPS` table are as follows: + +* `NAME`: the name of the resource group. +* `RU_PER_SEC`：the backfilling speed of the resource group in [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru)/second. +* `RU_TOKENS`: the number of tokens left in the resource group token bucket, and 1 token is an RU. +* `BURSTABLE`: whether to allow this resource group to overuse the remaining system resources. diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 37d84d0edefbc..2bd9ae3e18457 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -11,10 +11,10 @@ The `ALTER RESOURCE GROUP` statement is used to modify the resource group in a d ```ebnf+diagram AlterResourceGroupStmt: - "ALTER" "RESOURCE" "GROUP" IfNotExists ResourceGroupName ResourceGroupOptionList BurstableOption + "ALTER" "RESOURCE" "GROUP" IfNotExists ResourceGroupName ResourceGroupOptionList IfNotExists ::= - ('IF' 'NOT' 'EXISTS')? + ('IF' 'EXISTS')? ResourceGroupName: Identifier @@ -25,11 +25,8 @@ ResourceGroupOptionList: | ResourceGroupOptionList ',' DirectResourceGroupOption DirectResourceGroupOption: - "RRU_PER_SEC" EqOpt stringLit -| "WRU_PER_SEC" EqOpt stringLit - -BurstableOption ::= - ("BURSTABLE")? + "RU_PER_SEC" EqOpt stringLit +| "BURSTABLE" ``` @@ -39,7 +36,7 @@ TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU |---------------|-------------------------------------|------------------------| |`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RU per second | -If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota. +If the `BURSTABLE` attribute is set, the corresponding resource group allows the available system resources to be used exceeding the quota. > **Note:** > diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index eb67b8fa45ea5..277e83670eff3 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -11,7 +11,7 @@ You can use the `CREATE RESOURCE GROUP` statement to create a resource group in ```ebnf+diagram CreateResourceGroupStmt: - "CREATE" "RESOURCE" "GROUP" IfNotExists ResourceGroupName ResourceGroupOptionList BurstableOption + "CREATE" "RESOURCE" "GROUP" IfNotExists ResourceGroupName ResourceGroupOptionList IfNotExists ::= ('IF' 'NOT' 'EXISTS')? @@ -26,9 +26,7 @@ ResourceGroupOptionList: DirectResourceGroupOption: "RU_PER_SEC" EqOpt stringLit - -BurstableOption ::= - ("BURSTABLE")? +| "BURSTABLE" ``` @@ -59,7 +57,6 @@ mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 Query OK, 0 rows affected (0.08 sec) mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg2 -> RU_PER_SEC = 200 - -> BURSTABLE; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1' or NAME = 'rg2'; +------+-------------+-----------+-----------+ diff --git a/tidb-resource-control.md b/tidb-resource-control.md index ea3968409e510..4fe87a544a0f3 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -11,7 +11,14 @@ summary: Learn how to use resource management to control and schedule applicatio Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas by resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwhile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. -The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. +The TiDB resource control feature provides two layers of resource management capabilities: flow control capabilities at the TiDB layer and priority scheduling capabilities at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. + +- TiDB flow control: TiDB flow control uses the token bucket algorithm. The number of tokens consumed by read and write requests cannot exceed the number of tokens accumulated in the corresponding resource group bucket. If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry may fail due to timeout. +- TiKV scheduling: If this feature is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. + +The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. + +## Scenarios for resource control This feature applies to the following scenarios: From 32d8ee98c940357fbca5543b6aadc78c00908537 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Mon, 6 Feb 2023 22:49:10 +0800 Subject: [PATCH 21/63] Apply suggestions from code review MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Co-authored-by: Daniël van Eeden Co-authored-by: tiancaiamao --- information-schema/information-schema-processlist.md | 2 +- tidb-resource-control.md | 6 +++--- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/information-schema/information-schema-processlist.md b/information-schema/information-schema-processlist.md index 510d056bdebf8..05724f5946b66 100644 --- a/information-schema/information-schema-processlist.md +++ b/information-schema/information-schema-processlist.md @@ -38,7 +38,7 @@ DESC processlist; | MEM | bigint(21) unsigned | YES | | NULL | | | DISK | bigint(21) unsigned | YES | | NULL | | | TxnStart | varchar(64) | NO | | | | -| RESOURCE_GROUP | varchar(32) | NO | | | | +| RESOURCE_GROUP | varchar(32) | NO | | | | +---------------------+---------------------+------+------+---------+-------+ 13 rows in set (0.00 sec) ``` diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 4fe87a544a0f3..eb148612c79fb 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -9,7 +9,7 @@ summary: Learn how to use resource management to control and schedule applicatio > > This feature is experimental and its form and usage may change in subsequent versions. -Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas by resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwhile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. +Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas of resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwhile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. The TiDB resource control feature provides two layers of resource management capabilities: flow control capabilities at the TiDB layer and priority scheduling capabilities at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. @@ -31,7 +31,7 @@ In addition, the rational use of the resource control feature can reduce the num > **Note:** > -> This feature does not support [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). +> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). @@ -64,7 +64,7 @@ The resource control feature introduces two new global variables. > **Note:** > -> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. You need to restart the TiKV instance for the modification to take effect. +> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support](/tidb-cloud/tidb-cloud-support.md) to enable it. You need to restart the TiKV instance for the modification to take effect. From 5dcfb466cf7636ac7ec65ee91568001bba815b5b Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Tue, 7 Feb 2023 09:25:16 +0800 Subject: [PATCH 22/63] Update TOC.md --- TOC.md | 1 + 1 file changed, 1 insertion(+) diff --git a/TOC.md b/TOC.md index 97d2d7dbf80f7..25f67af6e744c 100644 --- a/TOC.md +++ b/TOC.md @@ -609,6 +609,7 @@ - [TiKV](/grafana-tikv-dashboard.md) - [TiFlash](/tiflash/monitor-tiflash.md) - [TiCDC](/ticdc/monitor-ticdc.md) + - [Resource Control](/grafana-resource-control-dashboard.md) - Secure - [Enable TLS Between TiDB Clients and Servers](/enable-tls-between-clients-and-servers.md) - [Enable TLS Between TiDB Components](/enable-tls-between-components.md) From 37ae62269edb0ae60c3d8add450e865811d2bc82 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Tue, 7 Feb 2023 14:56:22 +0800 Subject: [PATCH 23/63] polished wording --- grafana-resource-control-dashboard.md | 46 +++++++------- .../information-schema-resource-groups.md | 6 +- .../sql-statement-alter-resource-group.md | 2 +- .../sql-statement-create-resource-group.md | 2 +- .../sql-statement-show-privileges.md | 2 +- tidb-resource-control.md | 62 ++++++++++--------- 6 files changed, 63 insertions(+), 57 deletions(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index 16dd982daaf43..9b36744751250 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -11,26 +11,26 @@ The Grafana dashboard is divided into a series of sub dashboards which include O You can get an overview of the resource control status from the PD dashboard, where the key metrics are displayed. This document provides a detailed description of these key metrics. -The following is the description of Resource Control Dashboard metrics items. - -## Request Unit - -- RU: the [Request Unit](/tidb-resource-control.md#what-is-request-unit-ru) consumption information of each Resource Group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each Resource Group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). -- RU Per Query: the average number of Request Units consumed by each SQL statement per second is obtained by dividing the above Request Unit by the number of SQL statements executed per second. -- RRU: the Read Request Unit consumption information of each Resource Group is calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. -- RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second is obtained by dividing the above Read Request Unit by the number of SQL statements executed per second. -- WRU: the Write Request Unit consumption information of each Resource Group is calculated in real time. `total` is the sum of the Write Request Units consumed by all Resource Groups. -- WRU Per Query: the average number of Write Request Units consumed by each SQL statement per second is obtained by dividing the above Write Request Unit by the number of SQL statements executed per second. - -## Resource - -- KV Request Count: the number of KV requests is calculated in real time for each Resource Group. They are divided into read and write types. `total` is the sum of the KV requests involved by all Resource Groups. -- KV Request Count Per Query: the average number of read and write KV requests by each SQL statement per second is obtained by dividing the above KV Request Count by the number of SQL statements executed per second. -- Bytes Read: the amount of data read is calculated in real time for each Resource Group. `total` is the sum of the data read by all Resource Groups. -- Bytes Read Per Query: the average amount of data read by each SQL statement per second is obtained by dividing the above Bytes Read by the number of SQL statements executed per second. -- Bytes Written: the amount of data written is calculated in real time for each Resource Group. `total` is the sum of the data written by all Resource Groups. -- Bytes Written Per Query: the average amount of data written by each SQL statement per second is obtained by dividing the above Bytes Written by the number of SQL statements executed per second. -- KV CPU Time: the KV layer CPU time consumed is calculated in real time for each Resource Group. `total` is the sum of the KV layer CPU time consumed by all Resource Groups. -- KV CPU Time Per Query: the average KV layer CPU time consumed by each SQL statement per second is obtained by dividing the above KV CPU Time by the number of SQL statements executed per second. -- SQL CPU Time: the SQL layer CPU time consumed is calculated in real time for each Resource Group. `total` is the sum of the SQL layer CPU time consumed by all Resource Groups. -- SQL CPU Time Per Query: the average SQL layer CPU time consumed by each SQL statement per second is obtained by dividing the above SQL CPU Time by the number of SQL statements executed per second. +This document describes some key monitoring metrics displayed on the Resource Control Dashboard. + +## Metrics about Request Unit + +- RU: the [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) consumption information of each resource group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each resource group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). +- RU Per Query: the average number of Request Units consumed by each SQL statement per second. It is obtained by dividing the above Request Unit by the number of SQL statements executed per second. +- RRU: the Read Request Unit consumption information of each resource group is calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. +- RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second. It is obtained by dividing the above Read Request Unit by the number of SQL statements executed per second. +- WRU: the Write Request Unit consumption information of each resource group is calculated in real time. `total` is the sum of the Write Request Units consumed by all Resource Groups. +- WRU Per Query: the average number of Write Request Units consumed by each SQL statement per second. It is obtained by dividing the above Write Request Unit by the number of SQL statements executed per second. + +## Metrics about resources + +- KV Request Count: the number of KV requests is calculated in real time for each resource group. They are divided into read and write types. `total` is the sum of the KV requests involved by all Resource Groups. +- KV Request Count Per Query: the average number of read and write KV requests by each SQL statement per second. It is obtained by dividing the above KV Request Count by the number of SQL statements executed per second. +- Bytes Read: the amount of data read is calculated in real time for each resource group. `total` is the sum of the data read by all Resource Groups. +- Bytes Read Per Query: the average amount of data read by each SQL statement per second. It is obtained by dividing the above Bytes Read by the number of SQL statements executed per second. +- Bytes Written: the amount of data written is calculated in real time for each resource group. `total` is the sum of the data written by all Resource Groups. +- Bytes Written Per Query: the average amount of data written by each SQL statement per second. It is obtained by dividing the above Bytes Written by the number of SQL statements executed per second. +- KV CPU Time: the KV layer CPU time consumed is calculated in real time for each resource group. `total` is the sum of the KV layer CPU time consumed by all Resource Groups. +- KV CPU Time Per Query: the average KV layer CPU time consumed by each SQL statement per second. It is obtained by dividing the above KV CPU Time by the number of SQL statements executed per second. +- SQL CPU Time: the SQL layer CPU time consumed is calculated in real time for each resource group. `total` is the sum of the SQL layer CPU time consumed by all Resource Groups. +- SQL CPU Time Per Query: the average SQL layer CPU time consumed by each SQL statement per second. It is obtained by dividing the above SQL CPU Time by the number of SQL statements executed per second. diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index 165a09ccae692..72f404333020e 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -52,6 +52,6 @@ mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg1'; The descriptions of the columns in the `RESOURCE_GROUPS` table are as follows: * `NAME`: the name of the resource group. -* `RU_PER_SEC`：the backfilling speed of the resource group in [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru)/second. -* `RU_TOKENS`: the number of tokens left in the resource group token bucket, and 1 token is an RU. -* `BURSTABLE`: whether to allow this resource group to overuse the remaining system resources. +* `RU_PER_SEC`：the backfilling speed of the resource group. The unit is RU/second, in which RU means [Request Unit](/tidb-resource-control.md#what-is-request-unit-ru). +* `RU_TOKENS`: the number of tokens left in the resource group token bucket, and 1 token is 1 RU. +* `BURSTABLE`: whether to allow the resource group to overuse the available system resources. diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 2bd9ae3e18457..98f448672cf19 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -34,7 +34,7 @@ TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU | Option | Description | Example | |---------------|-------------------------------------|------------------------| -|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RU per second | +|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RUs per second | If the `BURSTABLE` attribute is set, the corresponding resource group allows the available system resources to be used exceeding the quota. diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index 277e83670eff3..1620a7436ee53 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -36,7 +36,7 @@ TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU | Option | Description | Example | |---------------|-------------------------------------|------------------------| -|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RU per second | +|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RUs per second | If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota. diff --git a/sql-statements/sql-statement-show-privileges.md b/sql-statements/sql-statement-show-privileges.md index 8b3852aaeb607..86ed22f609267 100644 --- a/sql-statements/sql-statement-show-privileges.md +++ b/sql-statements/sql-statement-show-privileges.md @@ -15,7 +15,7 @@ This statement shows a list of assignable privileges in TiDB. It is a static lis ## Examples -``` +```sql mysql> show privileges; +---------------------------------+---------------------------------------+-------------------------------------------------------+ | Privilege | Context | Comment | diff --git a/tidb-resource-control.md b/tidb-resource-control.md index eb148612c79fb..708672189028b 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -1,6 +1,6 @@ --- title: Use Resource Control to Achieve Resource Isolation -summary: Learn how to use resource management to control and schedule application resource consumption. +summary: Learn how to use the resource control feature to control and schedule application resources. --- # Use Resource Control to Achieve Resource Isolation @@ -11,13 +11,21 @@ summary: Learn how to use resource management to control and schedule applicatio Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas of resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwhile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. -The TiDB resource control feature provides two layers of resource management capabilities: flow control capabilities at the TiDB layer and priority scheduling capabilities at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. +The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. -- TiDB flow control: TiDB flow control uses the token bucket algorithm. The number of tokens consumed by read and write requests cannot exceed the number of tokens accumulated in the corresponding resource group bucket. If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry may fail due to timeout. -- TiKV scheduling: If this feature is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. +- TiDB flow control: TiDB flow control uses the token bucket algorithm (`RU_TOKENS`). The number of tokens consumed by read and write requests cannot exceed the number of tokens accumulated in the corresponding resource group bucket. If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry may fail due to timeout. +- TiKV scheduling: If [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. + + +> **Note:** +> +> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). + + + ## Scenarios for resource control This feature applies to the following scenarios: @@ -27,31 +35,23 @@ This feature applies to the following scenarios: In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. - - -> **Note:** -> -> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). - - - ## What is Request Unit (RU) -Request Unit (RU) is TiDB's unified abstraction unit for CPU, IO and other system resources, which currently includes CPU, IOPS and IO bandwidth metrics. The consumption of these three metrics is represented by the RU unit according to a certain ratio. +Request Unit (RU) is TiDB's unified abstraction unit for system resources, which currently includes CPU, IOPS and IO bandwidth metrics. The consumption of these three metrics is represented by RU according to a certain ratio. The following table shows the consumption of TiKV storage layer CPU and IO resources by user requests and the corresponding RU weights. -| Resource | RU Weight | -|:----------|:------| -| CPU | 1 RU / Milliseconds | -| Read IO | 1 RU / MiB | -| Write IO | 5 RU / MiB | -| Basic overhead of a read request | 1 RU | -| Basic overhead of a write request | 3 RU | +| Resource | RU Weight | +|:----------------|:-----------------| +| CPU | 1 RU/millisecond | +| Read IO | 1/64 RU/KB | +| Write IO | 1 RU/KB | +| Basic overhead of a read request | 0.25 RU | +| Basic overhead of a write request | 1.5 RU | -Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` MiB data, and `w1` write requests that write `w2` MiB data, then the formula for the total RUs consumed by the resource group is as follows: +Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, and `w1` write requests that write `w2` KB data, then the formula for the total RUs consumed by the resource group is as follows: -c + (r1 + r2) + (3 * w1 + 5 * w2) +`c` + (`r1` \* 0.25 + `r2`/64) + (1.5 * `w1` + `w2`) ## Parameters for resource control @@ -64,7 +64,7 @@ The resource control feature introduces two new global variables. > **Note:** > -> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support](/tidb-cloud/tidb-cloud-support.md) to enable it. You need to restart the TiKV instance for the modification to take effect. +> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. @@ -92,29 +92,35 @@ You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements ### Step 1. Enable the resource control feature -Enable flow control for the resource group. +Enable the resource control feature. ```sql SET GLOBAL tidb_enable_resource_control = 'ON'; ``` + + +In TiKV, set the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to `true`. + + + -In TiKV, set the parameter `resource_control.enabled` to `true`. The parameter `resource_control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. +In TiKV, set the parameter [`resource_control.enabled`](https://docs.pingcap.com/tidb/stable/tikv-configuration-file#resource_control) to `true`. The parameter `resource_control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. - + ### Step 2. Create a resource group, and then bind users to it The following is an example of how to create a resource group and bind users to it. -1. Create a resource group `rg1`. The RU backfill rate is 500 RU per second and allows applications in this resource group to overrun resources. +1. Create a resource group `rg1`. The RU backfill rate is 500 RUs per second and allows applications in this resource group to overrun resources. ```sql CREATE RESOURCE GROUP IF NOT EXISTS rg1 RU_PER_SEC = 500 BURSTABLE; ``` -2. Create a resource group `rg2`. The RU backfill rate is 500 RU per second and does not allow applications in this resource group to overrun resources. +2. Create a resource group `rg2`. The RU backfill rate is 500 RUs per second and does not allow applications in this resource group to overrun resources. ```sql CREATE RESOURCE GROUP IF NOT EXISTS rg2 RU_PER_SEC = 600; From ac5be5009afdb48c077dc016825b44b38004d246 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Tue, 7 Feb 2023 15:40:43 +0800 Subject: [PATCH 24/63] Update tidb-resource-control.md --- tidb-resource-control.md | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 708672189028b..cccf35eb732b4 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -142,6 +142,14 @@ If the resource group corresponds to a request with insufficient quota, the clie ## Monitoring and charts + + +> **Note:** +> +> This section is not applicable to TiDB Cloud. + + + TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** panel. The metrics are detailed in the **Resource Control** section of [TiDB Important Monitoring Metrics](/grafana-tidb-dashboard.md). ## Tool Compatibility From 8d4b2b5b28842db0a0d185cb41c79b4370ee8561 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Tue, 7 Feb 2023 16:31:41 +0800 Subject: [PATCH 25/63] Update information-schema-resource-groups.md --- information-schema/information-schema-resource-groups.md | 1 + 1 file changed, 1 insertion(+) diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index 72f404333020e..b52fd4281db88 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -48,6 +48,7 @@ mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg1'; +------+------------+-----------+-----------+ | rg1 | 1000 | 100000 | NO | +------+------------+-----------+-----------+ +``` The descriptions of the columns in the `RESOURCE_GROUPS` table are as follows: From 1a616a9bad0f3db1f11288bbfee388146447d19b Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Tue, 7 Feb 2023 18:37:09 +0800 Subject: [PATCH 26/63] Apply suggestions from code review Co-authored-by: BornChanger <97348524+BornChanger@users.noreply.github.com> Co-authored-by: ShuNing --- .../sql-statement-create-resource-group.md | 4 ++-- .../sql-statement-drop-resource-group.md | 2 +- tidb-resource-control.md | 18 +++++++++--------- tikv-configuration-file.md | 2 +- 4 files changed, 13 insertions(+), 13 deletions(-) diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index 1620a7436ee53..1a22181d22d88 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -5,7 +5,7 @@ summary: Learn the usage of CREATE RESOURCE GROUP in TiDB. # CREATE RESOURCE GROUP -You can use the `CREATE RESOURCE GROUP` statement to create a resource group in the currently selected database. +You can use the `CREATE RESOURCE GROUP` statement to create a resource group. ## Synopsis @@ -56,7 +56,7 @@ mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 -> BURSTABLE; Query OK, 0 rows affected (0.08 sec) mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg2 - -> RU_PER_SEC = 200 + -> RU_PER_SEC = 200; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1' or NAME = 'rg2'; +------+-------------+-----------+-----------+ diff --git a/sql-statements/sql-statement-drop-resource-group.md b/sql-statements/sql-statement-drop-resource-group.md index ae22d67263394..605d0c0accf1c 100644 --- a/sql-statements/sql-statement-drop-resource-group.md +++ b/sql-statements/sql-statement-drop-resource-group.md @@ -5,7 +5,7 @@ summary: Learn the usage of DROP RESOURCE GROUP in TiDB. # DROP RESOURCE GROUP -You can use the `DROP RESOURCE GROUP` statement to drop a resource group in the currently selected database. +You can use the `DROP RESOURCE GROUP` statement to drop a resource group. ## Synopsis diff --git a/tidb-resource-control.md b/tidb-resource-control.md index cccf35eb732b4..c782747323d5e 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -14,7 +14,7 @@ Using the resource control feature, as a cluster administrator, you can define r The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. - TiDB flow control: TiDB flow control uses the token bucket algorithm (`RU_TOKENS`). The number of tokens consumed by read and write requests cannot exceed the number of tokens accumulated in the corresponding resource group bucket. If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry may fail due to timeout. -- TiKV scheduling: If [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. +- TiKV scheduling: If [`resource-control.enabled`](/tikv-configuration-file.md#resource_control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. @@ -44,7 +44,7 @@ The following table shows the consumption of TiKV storage layer CPU and IO resou | Resource | RU Weight | |:----------------|:-----------------| | CPU | 1 RU/millisecond | -| Read IO | 1/64 RU/KB | +| Read IO | (1/64) RU/KB | | Write IO | 1 RU/KB | | Basic overhead of a read request | 0.25 RU | | Basic overhead of a write request | 1.5 RU | @@ -58,22 +58,22 @@ Based on the above table, assuming that the TiKV time consumed by a resource gro The resource control feature introduces two new global variables. * TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) to control whether to enable flow control for a resource group. -* TiKV: you can use the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. +* TiKV: you can use the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. > **Note:** > -> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. +> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it for TiDB Cloud Dedicated Tier clusters. The results of the combinations of these two parameters are shown in the following table. -| `resource_control.enabled` | `tidb_enable_resource_control`= ON | `tidb_enable_resource_control`= OFF | +| `resource-control.enabled` | `tidb_enable_resource_control`= ON | `tidb_enable_resource_control`= OFF | |:----------------------------|:-------------------------------------|:-------------------------------------| -| `resource_control.enabled`= true | Flow control and scheduling (recommended) | Invalid combination | -| `resource_control.enabled`= false | Only flow control | The feature is disabled. | +| `resource-control.enabled`= true | Flow control and scheduling (recommended) | Invalid combination | +| `resource-control.enabled`= false | Only flow control | The feature is disabled. | ## How to use resource control @@ -100,13 +100,13 @@ SET GLOBAL tidb_enable_resource_control = 'ON'; -In TiKV, set the parameter [`resource_control.enabled`](/tikv-configuration-file.md#resource_control) to `true`. +In TiKV, set the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource_control) to `true`. -In TiKV, set the parameter [`resource_control.enabled`](https://docs.pingcap.com/tidb/stable/tikv-configuration-file#resource_control) to `true`. The parameter `resource_control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. +In TiKV, set the parameter [`resource_control.enabled`](https://docs.pingcap.com/tidb/stable/tikv-configuration-file#resource_control) to `true`. The parameter `resource-control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. diff --git a/tikv-configuration-file.md b/tikv-configuration-file.md index 4e110b44a03c9..8681a773b04f6 100644 --- a/tikv-configuration-file.md +++ b/tikv-configuration-file.md @@ -2154,7 +2154,7 @@ To reduce write latency, TiKV periodically fetches and caches a batch of timesta + In a default TSO physical time update interval (`50ms`), PD provides at most 262144 TSOs. When requested TSOs exceed this number, PD provides no more TSOs. This configuration item is used to avoid exhausting TSOs and the reverse impact of TSO exhaustion on other businesses. If you increase the value of this configuration item to improve high availability, you need to decrease the value of [`tso-update-physical-interval`](/pd-configuration-file.md#tso-update-physical-interval) at the same time to get enough TSOs. + Default value: `8192` -## resource_control +## resource-control Configuration items related to resource control related to the TiKV storage layer. From da0538846c0a517e19c013e2b11f7119f6d6ecae Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Tue, 7 Feb 2023 21:55:51 +0800 Subject: [PATCH 27/63] Apply suggestions from code review --- tidb-resource-control.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index c782747323d5e..c63441c659203 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -64,7 +64,7 @@ The resource control feature introduces two new global variables. > **Note:** > -> The parameter `resource_control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it for TiDB Cloud Dedicated Tier clusters. +> The parameter `resource-control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it for TiDB Cloud Dedicated Tier clusters. @@ -106,7 +106,7 @@ In TiKV, set the parameter [`resource-control.enabled`](/tikv-configuration-file -In TiKV, set the parameter [`resource_control.enabled`](https://docs.pingcap.com/tidb/stable/tikv-configuration-file#resource_control) to `true`. The parameter `resource-control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. +In TiKV, set the parameter [`resource-control.enabled`](https://docs.pingcap.com/tidb/stable/tikv-configuration-file#resource_control) to `true`. The parameter `resource-control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. From bf8fd7e085226346c48141a5a1e66fbf89c4da93 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Tue, 7 Feb 2023 21:59:13 +0800 Subject: [PATCH 28/63] Apply suggestions from code review --- tidb-resource-control.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index c63441c659203..9a6d1f1b69eb5 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -14,7 +14,7 @@ Using the resource control feature, as a cluster administrator, you can define r The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. - TiDB flow control: TiDB flow control uses the token bucket algorithm (`RU_TOKENS`). The number of tokens consumed by read and write requests cannot exceed the number of tokens accumulated in the corresponding resource group bucket. If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry may fail due to timeout. -- TiKV scheduling: If [`resource-control.enabled`](/tikv-configuration-file.md#resource_control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. +- TiKV scheduling: If [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. @@ -58,7 +58,7 @@ Based on the above table, assuming that the TiKV time consumed by a resource gro The resource control feature introduces two new global variables. * TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) to control whether to enable flow control for a resource group. -* TiKV: you can use the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource_control) to control whether to use request scheduling based on resource group quotas. +* TiKV: you can use the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) to control whether to use request scheduling based on resource group quotas. @@ -100,7 +100,7 @@ SET GLOBAL tidb_enable_resource_control = 'ON'; -In TiKV, set the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource_control) to `true`. +In TiKV, set the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) to `true`. From e524039ac86252e41522fffb6419b9f63db8fc3e Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 8 Feb 2023 09:10:46 +0800 Subject: [PATCH 29/63] Update system-variables.md --- system-variables.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/system-variables.md b/system-variables.md index 98081189f355a..a262b8ef0a17d 100644 --- a/system-variables.md +++ b/system-variables.md @@ -1834,7 +1834,7 @@ MPP is a distributed computing framework provided by the TiFlash engine, which a - Persists to cluster: Yes - Default value: `OFF` - Type: Boolean -- This variable is a switch for [the resource control feature](/tidb-resource-control.md). When this variable is set to `ON`, the cluster supports resource isolation of applications by resource group. +- This variable is a switch for [the resource control feature](/tidb-resource-control.md). When this variable is set to `ON`, the TiDB cluster supports resource isolation of applications based on resource groups. ### tidb_enable_reuse_chunk New in v6.4.0 From 727ea1318a145ea73275b6b9d250a6598e57c45f Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 8 Feb 2023 09:17:52 +0800 Subject: [PATCH 30/63] Update tikv-configuration-file.md --- tikv-configuration-file.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tikv-configuration-file.md b/tikv-configuration-file.md index 8681a773b04f6..577e279f1dc04 100644 --- a/tikv-configuration-file.md +++ b/tikv-configuration-file.md @@ -2160,6 +2160,6 @@ Configuration items related to resource control related to the TiKV storage laye ### `enabled` New in v6.6.0 -+ Whether to support prioritized scheduling for user foreground read/write requests according to the corresponding resource group quota. For information about TiDB resource groups and resource control, see [TiDB resource control](/tidb-resource-control.md) -+ It is only meaningful to enable this configuration item if [`tidb_enable_resource_control](/system-variables.md#tidb_enable_resource_control-new-in-v660) is enabled on TiDB side. When it is enabled, TiKV will use the priority queue to schedule the queued requests for user foreground read/write requests. The priority of scheduling is inversely related to the amount of resources already consumed by the resource group where the request is located, and positively related to the quota of the corresponding resource group. -+ Default value: `false`, which means to disable scheduling based on resource group quotas. \ No newline at end of file ++ Whether to support scheduling for user foreground read/write requests according to Request Unit (RU) of the corresponding resource group. For information about TiDB resource groups and resource control, see [TiDB resource control](/tidb-resource-control.md). ++ It is only meaningful to enable this configuration item if [`tidb_enable_resource_control](/system-variables.md#tidb_enable_resource_control-new-in-v660) is enabled on TiDB. When it is enabled, TiKV will use the priority queue to schedule the queued requests for user foreground read/write requests. The priority of scheduling is inversely related to the amount of resources already consumed by the resource group where the request is located, and positively related to the quota of the corresponding resource group. ++ Default value: `false`, which means to disable scheduling based on the RU of the resource group. \ No newline at end of file From 4c82b2fafdf0c603bed35e55b557c40680a931f2 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 8 Feb 2023 09:39:26 +0800 Subject: [PATCH 31/63] Update tidb-resource-control.md --- tidb-resource-control.md | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 9a6d1f1b69eb5..a2dd15a0fbeeb 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -28,10 +28,12 @@ The introduction of the resource control feature is a milestone for TiDB. It can ## Scenarios for resource control -This feature applies to the following scenarios: + With this feature, you can: -- You can combine multiple small and medium-sized applications from different systems into one TiDB cluster. If the load of an individual application grows larger, it does not affect the normal operation of other businesses. When the system load is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. -- You can choose to combine all test environments into a single cluster, or group the batch tasks that consume more resources into a single resource group. It can improve hardware utilization and reduce operating costs while ensuring that critical applications can still get the necessary resources. + - Combine multiple small and medium-sized applications from different systems into a single TiDB cluster. When the workload of an application grows larger, it does not affect the normal operation of other applications. When the system workload is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. + - Choose to combine all test environments into a single TiDB cluster, or group the batch tasks that consume more resources into a single resource group. It can improve hardware utilization and reduce operating costs while ensuring that critical applications can always get the necessary resources. + + In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. From 9544e179eb908c82785ed67240cd34c9d474ab08 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Wed, 8 Feb 2023 15:53:55 +0800 Subject: [PATCH 32/63] sync cn commits --- .../information-schema-resource-groups.md | 12 +++++----- mysql-compatibility.md | 8 +++++++ .../sql-statement-alter-resource-group.md | 22 +++++++++---------- .../sql-statement-create-resource-group.md | 14 ++++++------ .../sql-statement-drop-resource-group.md | 12 +++++----- tidb-resource-control.md | 14 ++++++------ 6 files changed, 44 insertions(+), 38 deletions(-) diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index b52fd4281db88..045f126fe8249 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -22,7 +22,6 @@ DESC resource_groups; +------------+-------------+------+------+---------+-------+ | NAME | varchar(32) | NO | | NULL | | | RU_PER_SEC | bigint(21) | YES | | NULL | | -| RU_TOKENS | bigint(21) | YES | | NULL | | | BURSTABLE | varchar(3) | YES | | NULL | | +------------+-------------+------+------+---------+-------+ 4 rows in set (0.00 sec) @@ -43,16 +42,15 @@ mysql> SHOW CREATE RESOURCE GROUP rg1; -- Displays the definition of the rg1 res +----------------+---------------------------------------------+ 1 row in set (0.00 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg1'; -+------+------------+-----------+-----------+ -| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | -+------+------------+-----------+-----------+ -| rg1 | 1000 | 100000 | NO | -+------+------------+-----------+-----------+ ++------+------------+-----------+ +| NAME | RU_PER_SEC | BURSTABLE | ++------+------------+-----------+ +| rg1 | 1000 | NO | ++------+------------+-----------+ ``` The descriptions of the columns in the `RESOURCE_GROUPS` table are as follows: * `NAME`: the name of the resource group. * `RU_PER_SEC`：the backfilling speed of the resource group. The unit is RU/second, in which RU means [Request Unit](/tidb-resource-control.md#what-is-request-unit-ru). -* `RU_TOKENS`: the number of tokens left in the resource group token bucket, and 1 token is 1 RU. * `BURSTABLE`: whether to allow the resource group to overuse the available system resources. diff --git a/mysql-compatibility.md b/mysql-compatibility.md index a0587f91322d9..cf901f016cc02 100644 --- a/mysql-compatibility.md +++ b/mysql-compatibility.md @@ -255,3 +255,11 @@ TiDB does not implement certain features that have been marked as deprecated in * Specifying precision for floating point types. MySQL 8.0 [deprecates](https://dev.mysql.com/doc/refman/8.0/en/floating-point-types.html) this feature, and it is recommended to use the `DECIMAL` type instead. * The `ZEROFILL` attribute. MySQL 8.0 [deprecates](https://dev.mysql.com/doc/refman/8.0/en/numeric-type-attributes.html) this feature, and it is recommended to instead pad numeric values in your application. + +### `CREATE RESOURCE GROUP`, `DROP RESOURCE GROUP`, and `ALTER RESOURCE GROUP` statements + +The statements of creating and modifying resource groups are different from MySQL. See the following documents for details: + +- [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md) +- [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md) +- [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md) diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 98f448672cf19..d77a897eac6b5 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -40,7 +40,7 @@ If the `BURSTABLE` attribute is set, the corresponding resource group allows the > **Note:** > -> The `ALTER RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_group`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. +> The `ALTER RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_control`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. ## Examples @@ -54,21 +54,21 @@ mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 -> BURSTABLE; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+-------------+-----------+-----------+ -| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | -+------+-------------+-----------+-----------+ -| rg1 | 100 | 165135 | YES | -+------+-------------+-----------+-----------+ ++------+-------------+-----------+ +| NAME | RU_PER_SEC | BURSTABLE | ++------+-------------+-----------+ +| rg1 | 100 | YES | ++------+-------------+-----------+ 1 rows in set (1.30 sec) mysql> ALTER RESOURCE GROUP rg1 -> RU_PER_SEC = 200; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+-------------+-----------+-----------+ -| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | -+------+-------------+-----------+-----------+ -| rg1 | 200 | 257158 | NO | -+------+-------------+-----------+-----------+ ++------+-------------+-----------+ +| NAME | RU_PER_SEC | BURSTABLE | ++------+-------------+-----------+ +| rg1 | 200 | NO | ++------+-------------+-----------+ 1 rows in set (1.30 sec) ``` diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index 1a22181d22d88..4c985404319e0 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -42,7 +42,7 @@ If the `BURSTABLE` attribute is set, the corresponding resource group allows the > **Note:** > -> The `CREATE RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_group`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. +> The `CREATE RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_control`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. ## Examples @@ -59,12 +59,12 @@ mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg2 -> RU_PER_SEC = 200; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1' or NAME = 'rg2'; -+------+-------------+-----------+-----------+ -| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | -+------+-------------+-----------+-----------+ -| rg1 | 100 | 165135 | YES | -| rg2 | 200 | 157158 | NO | -+------+-------------+-----------+-----------+ ++------+-------------+-----------+ +| NAME | RU_PER_SEC | BURSTABLE | ++------+-------------+-----------+ +| rg1 | 100 | YES | +| rg2 | 200 | NO | ++------+-------------+-----------+ 2 rows in set (1.30 sec) ``` diff --git a/sql-statements/sql-statement-drop-resource-group.md b/sql-statements/sql-statement-drop-resource-group.md index 605d0c0accf1c..1ca96cf098efe 100644 --- a/sql-statements/sql-statement-drop-resource-group.md +++ b/sql-statements/sql-statement-drop-resource-group.md @@ -22,7 +22,7 @@ ResourceGroupName: > **Note:** > -> The `DROP RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_group`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. +> The `DROP RESOURCE GROUP` statement can only be executed when the global variable [`tidb_enable_resource_control`](/system-variables.md#tidb_enable_resource_control-new-in-v660) is set to `ON`. ## Examples @@ -34,11 +34,11 @@ Query OK, 0 rows affected (0.22 sec) mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 RU_PER_SEC = 500 BURSTABLE; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+------------+-----------+-----------+ -| NAME | RU_PER_SEC | RU_TOKENS | BURSTABLE | -+------+------------+-----------+-----------+ -| rg1 | 500 | 0 | YES | -+------+------------+-----------+-----------+ ++------+------------+-----------+ +| NAME | RU_PER_SEC | BURSTABLE | ++------+------------+-----------+ +| rg1 | 500 | YES | ++------+------------+-----------+ 1 row in set (0.01 sec) mysql> DROP RESOURCE GROUP IF EXISTS rg1 ; diff --git a/tidb-resource-control.md b/tidb-resource-control.md index a2dd15a0fbeeb..b85c7669be0c8 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -28,12 +28,12 @@ The introduction of the resource control feature is a milestone for TiDB. It can ## Scenarios for resource control - With this feature, you can: +With this feature, you can: - - Combine multiple small and medium-sized applications from different systems into a single TiDB cluster. When the workload of an application grows larger, it does not affect the normal operation of other applications. When the system workload is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. - - Choose to combine all test environments into a single TiDB cluster, or group the batch tasks that consume more resources into a single resource group. It can improve hardware utilization and reduce operating costs while ensuring that critical applications can always get the necessary resources. +- Combine multiple small and medium-sized applications from different systems into a single TiDB cluster. When the workload of an application grows larger, it does not affect the normal operation of other applications. When the system workload is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. +- Choose to combine all test environments into a single TiDB cluster, or group the batch tasks that consume more resources into a single resource group. It can improve hardware utilization and reduce operating costs while ensuring that critical applications can always get the necessary resources. - In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. +In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. @@ -45,7 +45,7 @@ The following table shows the consumption of TiKV storage layer CPU and IO resou | Resource | RU Weight | |:----------------|:-----------------| -| CPU | 1 RU/millisecond | +| CPU | (1/3) RU/millisecond | | Read IO | (1/64) RU/KB | | Write IO | 1 RU/KB | | Basic overhead of a read request | 0.25 RU | @@ -53,7 +53,7 @@ The following table shows the consumption of TiKV storage layer CPU and IO resou Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, and `w1` write requests that write `w2` KB data, then the formula for the total RUs consumed by the resource group is as follows: -`c` + (`r1` \* 0.25 + `r2`/64) + (1.5 * `w1` + `w2`) +`c`/3 + (`r1` \* 0.25 + `r2`/64) + (1.5 * `w1` + `w2`) ## Parameters for resource control @@ -156,7 +156,7 @@ TiDB regularly collects runtime information about resource control and provides ## Tool Compatibility -The resource control feature is not compatible with data export/import and replication tools including BR, TiDB Lightning, and TiCDC. +The resource control feature is not compatible with data export/import and replication tools including BR, TiDB Lightning, and TiCDC. These tools simply ignore the metadata information of the resource groups in the cluster. ## Limitations From 5412337946cd0b25c0f6e1259b04c4b5fd2bb64e Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 8 Feb 2023 21:37:05 +0800 Subject: [PATCH 33/63] Apply suggestions from code review --- .../sql-statement-show-create-resource-group.md | 13 +++++++++---- system-variables.md | 2 +- tidb-resource-control.md | 6 +++--- 3 files changed, 13 insertions(+), 8 deletions(-) diff --git a/sql-statements/sql-statement-show-create-resource-group.md b/sql-statements/sql-statement-show-create-resource-group.md index e3a8ae72071c3..9b3cc33927338 100644 --- a/sql-statements/sql-statement-show-create-resource-group.md +++ b/sql-statements/sql-statement-show-create-resource-group.md @@ -19,19 +19,24 @@ ResourceGroupName ::= ## Examples -View the current definition of the resource group `rg1`. +Create a resource group `rg1`. ```sql CREATE RESOURCE GROUP rg1 RU_PER_SEC=100; Query OK, 0 rows affected (0.10 sec) ``` +View the definition of `rg1`. + ```sql SHOW CREATE RESOURCE GROUP rg1; ***************************[ 1. row ]*************************** -Resource_Group | rg1 -Create Resource Group | CREATE RESOURCE GROUP `rg1` RU_PER_SEC=100 -1 row in set (0.00 sec) ++----------------+--------------------------------------------+ +| Resource_Group | Create Resource Group | ++----------------+--------------------------------------------+ +| rg1 | CREATE RESOURCE GROUP `rg1` RU_PER_SEC=100 | ++----------------+--------------------------------------------+ +1 row in set (0.01 sec) ``` ## MySQL compatibility diff --git a/system-variables.md b/system-variables.md index a262b8ef0a17d..8731333dca38b 100644 --- a/system-variables.md +++ b/system-variables.md @@ -1834,7 +1834,7 @@ MPP is a distributed computing framework provided by the TiFlash engine, which a - Persists to cluster: Yes - Default value: `OFF` - Type: Boolean -- This variable is a switch for [the resource control feature](/tidb-resource-control.md). When this variable is set to `ON`, the TiDB cluster supports resource isolation of applications based on resource groups. +- This variable is a switch for [the resource control feature](/tidb-resource-control.md). When this variable is set to `ON`, the TiDB cluster can isolate application resources based on resource groups. ### tidb_enable_reuse_chunk New in v6.4.0 diff --git a/tidb-resource-control.md b/tidb-resource-control.md index b85c7669be0c8..5dc9a8551a9e9 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -9,11 +9,11 @@ summary: Learn how to use the resource control feature to control and schedule a > > This feature is experimental and its form and usage may change in subsequent versions. -Using the resource control feature, as a cluster administrator, you can define resource groups and limit the read and write quotas of resource groups. After you bind users to a resource group, the TiDB layer will perform flow control on the user's read and write requests according to the read and write quotas set by the bound resource group. Meanwhile, the TiKV layer will schedule the requests according to the priority of the read and write quota mapping. Through flow control and scheduling, you can achieve resource isolation of your applications and meet the quality of service (QoS) requirements. +As a cluster administrator, you can use the resource control feature to set read and write quotas for resource groups, and bind users to those groups. This allows the TiDB layer to control the flow of user read and write requests based on the quotas set for the resource group, while the TiKV layer schedules the requests based on the priority of the read and write quota mapping. By doing this, you can ensure resource isolation for your applications and meet quality of service (QoS) requirements. The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. -- TiDB flow control: TiDB flow control uses the token bucket algorithm (`RU_TOKENS`). The number of tokens consumed by read and write requests cannot exceed the number of tokens accumulated in the corresponding resource group bucket. If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry may fail due to timeout. +- TiDB flow control: TiDB flow control uses the [token bucket algorithm](https://en.wikipedia.org/wiki/Token_bucket). If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry might fail due to timeout. - TiKV scheduling: If [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. @@ -156,7 +156,7 @@ TiDB regularly collects runtime information about resource control and provides ## Tool Compatibility -The resource control feature is not compatible with data export/import and replication tools including BR, TiDB Lightning, and TiCDC. These tools simply ignore the metadata information of the resource groups in the cluster. +The resource control feature is still in its experimental stage and does not impact the regular usage of data import, export, and other replication tools. BR, TiDB Lightning, and TiCDC do not currently support resource control related DDL processing, and their resource consumption is not limited by resource control. ## Limitations From 9db7cbebd7f61a8691b224f9d7d57d093745a6dd Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 9 Feb 2023 17:21:02 +0800 Subject: [PATCH 34/63] Apply suggestions from code review --- tidb-resource-control.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 5dc9a8551a9e9..41378d7e8abcd 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -53,7 +53,7 @@ The following table shows the consumption of TiKV storage layer CPU and IO resou Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, and `w1` write requests that write `w2` KB data, then the formula for the total RUs consumed by the resource group is as follows: -`c`/3 + (`r1` \* 0.25 + `r2`/64) + (1.5 * `w1` + `w2`) +`c`\* 1/3 + (`r1` \* 0.25 + `r2` \* 1/64) + (1.5 \* `w1` + `w2` \* 1 \* `n`) ## Parameters for resource control @@ -75,7 +75,7 @@ The results of the combinations of these two parameters are shown in the followi | `resource-control.enabled` | `tidb_enable_resource_control`= ON | `tidb_enable_resource_control`= OFF | |:----------------------------|:-------------------------------------|:-------------------------------------| | `resource-control.enabled`= true | Flow control and scheduling (recommended) | Invalid combination | -| `resource-control.enabled`= false | Only flow control | The feature is disabled. | +| `resource-control.enabled`= false | Only flow control (not recommended) | The feature is disabled. | ## How to use resource control From b5f16d44e295508f7848755f1315830dff1fda50 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 9 Feb 2023 17:24:26 +0800 Subject: [PATCH 35/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 41378d7e8abcd..0db34ac7b59ac 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -51,7 +51,7 @@ The following table shows the consumption of TiKV storage layer CPU and IO resou | Basic overhead of a read request | 0.25 RU | | Basic overhead of a write request | 1.5 RU | -Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, and `w1` write requests that write `w2` KB data, then the formula for the total RUs consumed by the resource group is as follows: +Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, `w1` write requests that write `w2` KB data, and the number of non-witness TiKV nodes in the cluster is `n` , then the formula for the total RUs consumed by the resource group is as follows: `c`\* 1/3 + (`r1` \* 0.25 + `r2` \* 1/64) + (1.5 \* `w1` + `w2` \* 1 \* `n`) From 37518b485a8a00e13e6b3c581683086e71a5b738 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 9 Feb 2023 17:25:00 +0800 Subject: [PATCH 36/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 0db34ac7b59ac..f5fc29ddb3e56 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -51,7 +51,7 @@ The following table shows the consumption of TiKV storage layer CPU and IO resou | Basic overhead of a read request | 0.25 RU | | Basic overhead of a write request | 1.5 RU | -Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, `w1` write requests that write `w2` KB data, and the number of non-witness TiKV nodes in the cluster is `n` , then the formula for the total RUs consumed by the resource group is as follows: +Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, `w1` write requests that write `w2` KB data, and the number of non-witness TiKV nodes in the cluster is `n`, then the formula for the total RUs consumed by the resource group is as follows: `c`\* 1/3 + (`r1` \* 0.25 + `r2` \* 1/64) + (1.5 \* `w1` + `w2` \* 1 \* `n`) From 9009b81b70266e4da56049f6ec35d95e99efbddd Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 9 Feb 2023 21:06:02 +0800 Subject: [PATCH 37/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 -- 1 file changed, 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index f5fc29ddb3e56..eb6fcae93c0d4 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -35,8 +35,6 @@ With this feature, you can: In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. -In addition, the rational use of the resource control feature can reduce the number of clusters, ease the difficulty of operation and maintenance, and save management costs. - ## What is Request Unit (RU) Request Unit (RU) is TiDB's unified abstraction unit for system resources, which currently includes CPU, IOPS and IO bandwidth metrics. The consumption of these three metrics is represented by RU according to a certain ratio. From d998e1bf2359fe6ca0ab5cbbfe0fa844696830aa Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Sat, 11 Feb 2023 22:36:16 +0800 Subject: [PATCH 38/63] Apply suggestions from code review Co-authored-by: Grace Cai Co-authored-by: glorv --- grafana-resource-control-dashboard.md | 36 +++++++------- .../information-schema-resource-groups.md | 8 ++-- mysql-compatibility.md | 2 +- .../sql-statement-alter-resource-group.md | 30 ++++++------ .../sql-statement-create-resource-group.md | 10 ++-- sql-statements/sql-statement-create-user.md | 10 ++-- system-variables.md | 2 +- tidb-resource-control.md | 48 ++++++++++--------- tikv-configuration-file.md | 6 +-- 9 files changed, 77 insertions(+), 75 deletions(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index 9b36744751250..1413e64cf71b4 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -7,30 +7,30 @@ summary: Learn some key metrics displayed on the Grafana Resource Control dashbo If you use TiUP to deploy the TiDB cluster, the monitoring system (Prometheus & Grafana) is deployed at the same time. For more information, see [Overview of the Monitoring Framework](/tidb-monitoring-framework.md). -The Grafana dashboard is divided into a series of sub dashboards which include Overview, PD, TiDB, TiKV, Node\_exporter, Disk Performance, and Performance\_overview. A lot of metrics are there to help you diagnose. +The Grafana dashboard is divided into a series of sub dashboards which include Overview, PD, TiDB, TiKV, Node\_exporter, Disk Performance, and Performance\_overview. -You can get an overview of the resource control status from the PD dashboard, where the key metrics are displayed. This document provides a detailed description of these key metrics. +If your cluster has used the [Resource Control](tidb-resource-control.md) feature, you can get an overview of the resource consumption status from the Resource Control dashboard. -This document describes some key monitoring metrics displayed on the Resource Control Dashboard. +This document describes some key monitoring metrics displayed on the Resource Control dashboard. ## Metrics about Request Unit -- RU: the [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) consumption information of each resource group is calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each resource group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). -- RU Per Query: the average number of Request Units consumed by each SQL statement per second. It is obtained by dividing the above Request Unit by the number of SQL statements executed per second. -- RRU: the Read Request Unit consumption information of each resource group is calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. -- RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second. It is obtained by dividing the above Read Request Unit by the number of SQL statements executed per second. -- WRU: the Write Request Unit consumption information of each resource group is calculated in real time. `total` is the sum of the Write Request Units consumed by all Resource Groups. -- WRU Per Query: the average number of Write Request Units consumed by each SQL statement per second. It is obtained by dividing the above Write Request Unit by the number of SQL statements executed per second. +- RU: the [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) consumption information of each resource group, calculated in real time. `total` is the sum of the Request Units consumed by all Resource Groups. The Request Unit consumption of each resource group should be equal to the sum of its read consumption (Read Request Unit) and write consumption (Write Request Unit). +- RU Per Query: the average number of Request Units consumed by each SQL statement per second. It is obtained by dividing the above RU metric by the number of SQL statements executed per second. +- RRU: the Read Request Unit consumption information of each resource group, calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. +- RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second. It is obtained by dividing the above RRU metric by the number of SQL statements executed per second. +- WRU: the Write Request Unit consumption information of each resource group, calculated in real time. `total` is the sum of the Write Request Units consumed by all Resource Groups. +- WRU Per Query: the average number of Write Request Units consumed by each SQL statement per second. It is obtained by dividing the above WRU metric by the number of SQL statements executed per second. ## Metrics about resources - KV Request Count: the number of KV requests is calculated in real time for each resource group. They are divided into read and write types. `total` is the sum of the KV requests involved by all Resource Groups. -- KV Request Count Per Query: the average number of read and write KV requests by each SQL statement per second. It is obtained by dividing the above KV Request Count by the number of SQL statements executed per second. -- Bytes Read: the amount of data read is calculated in real time for each resource group. `total` is the sum of the data read by all Resource Groups. -- Bytes Read Per Query: the average amount of data read by each SQL statement per second. It is obtained by dividing the above Bytes Read by the number of SQL statements executed per second. -- Bytes Written: the amount of data written is calculated in real time for each resource group. `total` is the sum of the data written by all Resource Groups. -- Bytes Written Per Query: the average amount of data written by each SQL statement per second. It is obtained by dividing the above Bytes Written by the number of SQL statements executed per second. -- KV CPU Time: the KV layer CPU time consumed is calculated in real time for each resource group. `total` is the sum of the KV layer CPU time consumed by all Resource Groups. -- KV CPU Time Per Query: the average KV layer CPU time consumed by each SQL statement per second. It is obtained by dividing the above KV CPU Time by the number of SQL statements executed per second. -- SQL CPU Time: the SQL layer CPU time consumed is calculated in real time for each resource group. `total` is the sum of the SQL layer CPU time consumed by all Resource Groups. -- SQL CPU Time Per Query: the average SQL layer CPU time consumed by each SQL statement per second. It is obtained by dividing the above SQL CPU Time by the number of SQL statements executed per second. +- KV Request Count Per Query: the average number of read and write KV requests by each SQL statement per second. It is obtained by dividing the above KV Request Count metric by the number of SQL statements executed per second. +- Bytes Read: the amount of data read by each Resource Group, calculated per second. `total` is the sum of the data read by all Resource Groups. +- Bytes Read Per Query: the average amount of data read by each SQL statement per second. It is obtained by dividing the above Bytes Read metric by the number of SQL statements executed per second. +- Bytes Written: the amount of data written by each Resource Group, calculated in real time. `total` is the sum of the data written by all Resource Groups. +- Bytes Written Per Query: the average amount of data written by each SQL statement per second. It is obtained by dividing the above Bytes Written metric by the number of SQL statements executed per second. +- KV CPU Time: the KV layer CPU time consumed by each Resource Group, calculated in real time . `total` is the sum of the KV layer CPU time consumed by all Resource Groups. +- KV CPU Time Per Query: the average KV layer CPU time consumed by each SQL statement per second. It is obtained by dividing the above KV CPU Time metric by the number of SQL statements executed per second. +- SQL CPU Time: the SQL layer CPU time consumed by each Resource Group, calculated in real time. `total` is the sum of the SQL layer CPU time consumed by all Resource Groups. +- SQL CPU Time Per Query: the average SQL layer CPU time consumed by each SQL statement per second. It is obtained by dividing the above SQL CPU Time metric by the number of SQL statements executed per second. diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index 045f126fe8249..a2993c385a9d1 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -7,9 +7,9 @@ summary: Learn the `RESOURCE_GROUPS` information_schema table. > **Warning:** > -> This feature is experimental and its form and usage may change in subsequent versions. +> This feature is experimental and its form and usage might change in subsequent versions. -The `RESOURCE_GROUPS` table shows information about all resource groups. See [Use Resource Control to Achieve Resource Isolation](/tidb-resource-control.md). +The `RESOURCE_GROUPS` table shows the information about all resource groups. For more information, see [Use Resource Control to Achieve Resource Isolation](/tidb-resource-control.md). ```sql USE information_schema; @@ -29,12 +29,10 @@ DESC resource_groups; ## Examples -{{< copyable "sql" >}} - ```sql mysql> CREATE RESOURCE GROUP rg1 RU_PER_SEC=1000; -- Create the resource group rg1 Query OK, 0 rows affected (0.34 sec) -mysql> SHOW CREATE RESOURCE GROUP rg1; -- Displays the definition of the rg1 resource group. +mysql> SHOW CREATE RESOURCE GROUP rg1; -- Display the definition of the rg1 resource group +----------------+---------------------------------------------+ | Resource_Group | Create Resource Group | +----------------+---------------------------------------------+ diff --git a/mysql-compatibility.md b/mysql-compatibility.md index cf901f016cc02..28bbdd2ad20d7 100644 --- a/mysql-compatibility.md +++ b/mysql-compatibility.md @@ -258,7 +258,7 @@ TiDB does not implement certain features that have been marked as deprecated in ### `CREATE RESOURCE GROUP`, `DROP RESOURCE GROUP`, and `ALTER RESOURCE GROUP` statements -The statements of creating and modifying resource groups are different from MySQL. See the following documents for details: +For the statements of creating, modifying, and dropping resource groups, the supported parameters are different from that of MySQL. See the following documents for details: - [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md) - [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md) diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index d77a897eac6b5..56df867d26a66 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -5,7 +5,7 @@ summary: Learn the usage of ALTER RESOURCE GROUP in TiDB. # ALTER RESOURCE GROUP -The `ALTER RESOURCE GROUP` statement is used to modify the resource group in a database. +The `ALTER RESOURCE GROUP` statement is used to modify a resource group in a database. ## Synopsis @@ -30,13 +30,13 @@ DirectResourceGroupOption: ``` -TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) is TiDB's unified abstraction unit for CPU, IO, and other system resources. +TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) is a unified abstraction unit in TiDB for CPU, IO, and other system resources. | Option | Description | Example | |---------------|-------------------------------------|------------------------| -|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RUs per second | +| `RU_PER_SEC` | Rate of RU backfilling per second | `RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RUs per second | -If the `BURSTABLE` attribute is set, the corresponding resource group allows the available system resources to be used exceeding the quota. +If the `BURSTABLE` attribute is set, TiDB allows the corresponding resource group to use the available system resources when the quota is exceeded. > **Note:** > @@ -54,27 +54,27 @@ mysql> CREATE RESOURCE GROUP IF NOT EXISTS rg1 -> BURSTABLE; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+-------------+-----------+ -| NAME | RU_PER_SEC | BURSTABLE | -+------+-------------+-----------+ -| rg1 | 100 | YES | -+------+-------------+-----------+ ++------+------------+-----------+ +| NAME | RU_PER_SEC | BURSTABLE | ++------+------------+-----------+ +| rg1 | 100 | YES | ++------+------------+-----------+ 1 rows in set (1.30 sec) mysql> ALTER RESOURCE GROUP rg1 -> RU_PER_SEC = 200; Query OK, 0 rows affected (0.08 sec) mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; -+------+-------------+-----------+ -| NAME | RU_PER_SEC | BURSTABLE | -+------+-------------+-----------+ -| rg1 | 200 | NO | -+------+-------------+-----------+ ++------+------------+-----------+ +| NAME | RU_PER_SEC | BURSTABLE | ++------+------------+-----------+ +| rg1 | 200 | NO | ++------+------------+-----------+ 1 rows in set (1.30 sec) ``` ## MySQL compatibility -MySQL also supports [ALTER RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/en/alter-resource-group.html). However, the accepted parameters are different from TiDB so that they are not compatible. +MySQL also supports [ALTER RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/en/alter-resource-group.html). However, the acceptable parameters are different from that of TiDB so that they are not compatible. ## See also diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index 4c985404319e0..c721ff2fd5009 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -30,15 +30,15 @@ DirectResourceGroupOption: ``` -The resource group name parameter (`ResourceGroupName`) is globally unique and cannot be duplicated. +The resource group name parameter (`ResourceGroupName`) must be globally unique. -TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) is TiDB's unified abstraction unit for CPU, IO, and other system resources. +TiDB supports the following `DirectResourceGroupOption`, where [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) is a unified abstraction unit in TiDB for CPU, IO, and other system resources. | Option | Description | Example | |---------------|-------------------------------------|------------------------| -|`RU_PER_SEC` | Rate of RU filling per second |`RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RUs per second | +| `RU_PER_SEC` | Rate of RU backfilling per second | `RU_PER_SEC = 500` indicates that this resource group is backfilled with 500 RUs per second | -If the `BURSTABLE` attribute is set, the corresponding resource group allows the system resources to be used exceeding the quota. +If the `BURSTABLE` attribute is set, TiDB allows the corresponding resource group to use the available system resources when the quota is exceeded. > **Note:** > @@ -70,7 +70,7 @@ mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1' or NAM ## MySQL compatibility -MySQL also supports [CREATE RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/en/create-resource-group.html). However, the accepted parameters are different from TiDB so that they are not compatible. +MySQL also supports [CREATE RESOURCE GROUP](https://dev.mysql.com/doc/refman/8.0/en/create-resource-group.html). However, the acceptable parameters are different from that of TiDB so that they are not compatible. ## See also diff --git a/sql-statements/sql-statement-create-user.md b/sql-statements/sql-statement-create-user.md index 607240c59957c..91c0bf61a193a 100644 --- a/sql-statements/sql-statement-create-user.md +++ b/sql-statements/sql-statement-create-user.md @@ -138,11 +138,11 @@ SELECT USER, HOST, USER_ATTRIBUTES FROM MYSQL.USER WHERE USER='newuser7'; ``` ```sql -+-----------+------+---------------------------------------------------+ -| USER | HOST | ATTRIBUTE | -+-----------+------+---------------------------------------------------+ -| newuser7 | % | {"resource_group": "rg1"} | -+-----------+------+---------------------------------------------------+ ++----------+------+---------------------------+ +| USER | HOST | USER_ATTRIBUTES | ++----------+------+---------------------------+ +| newuser7 | % | {"resource_group": "rg1"} | ++----------+------+---------------------------+ 1 rows in set (0.00 sec) ``` diff --git a/system-variables.md b/system-variables.md index 8731333dca38b..bda6c30c300dd 100644 --- a/system-variables.md +++ b/system-variables.md @@ -1828,7 +1828,7 @@ MPP is a distributed computing framework provided by the TiFlash engine, which a > **Warning:** > -> This feature is experimental and its form and usage may change in subsequent versions. +> This feature is experimental and its form and usage might change in subsequent versions. - Scope: GLOBAL - Persists to cluster: Yes diff --git a/tidb-resource-control.md b/tidb-resource-control.md index eb6fcae93c0d4..dc709a535f14e 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -7,14 +7,14 @@ summary: Learn how to use the resource control feature to control and schedule a > **Warning:** > -> This feature is experimental and its form and usage may change in subsequent versions. +> This feature is experimental and its form and usage might change in subsequent versions. -As a cluster administrator, you can use the resource control feature to set read and write quotas for resource groups, and bind users to those groups. This allows the TiDB layer to control the flow of user read and write requests based on the quotas set for the resource group, while the TiKV layer schedules the requests based on the priority of the read and write quota mapping. By doing this, you can ensure resource isolation for your applications and meet quality of service (QoS) requirements. +As a cluster administrator, you can use the resource control feature to create resource groups, set read and write quotas for resource groups, and bind users to those groups. This allows the TiDB layer to control the flow of user read and write requests based on the quotas set for the resource groups, and allows the TiKV layer to schedule the requests based on the priority mapped to the read and write quota. By doing this, you can ensure resource isolation for your applications and meet quality of service (QoS) requirements. -The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities are orthogonal and can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. +The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. -- TiDB flow control: TiDB flow control uses the [token bucket algorithm](https://en.wikipedia.org/wiki/Token_bucket). If there are not enough tokens in the bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry might fail due to timeout. -- TiKV scheduling: If [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to map the read and write requests of each resource group to their own priorities. Based on their own priorities, the storage layer uses the priority queue to schedule and process requests. +- TiDB flow control: TiDB flow control uses the [token bucket algorithm](https://en.wikipedia.org/wiki/Token_bucket). If there are not enough tokens in a bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry might fail due to timeout. +- TiKV scheduling: if [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to determine the priority of the read and write requests for each resource group. Based on the priorities, the storage layer uses the priority queue to schedule and process requests. The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. @@ -28,6 +28,8 @@ The introduction of the resource control feature is a milestone for TiDB. It can ## Scenarios for resource control +The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. + With this feature, you can: - Combine multiple small and medium-sized applications from different systems into a single TiDB cluster. When the workload of an application grows larger, it does not affect the normal operation of other applications. When the system workload is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. @@ -37,19 +39,19 @@ In addition, the rational use of the resource control feature can reduce the num ## What is Request Unit (RU) -Request Unit (RU) is TiDB's unified abstraction unit for system resources, which currently includes CPU, IOPS and IO bandwidth metrics. The consumption of these three metrics is represented by RU according to a certain ratio. +Request Unit (RU) is a unified abstraction unit in TiDB for system resources, which currently includes CPU, IOPS, and IO bandwidth metrics. The consumption of these three metrics is represented by RU according to a certain ratio. The following table shows the consumption of TiKV storage layer CPU and IO resources by user requests and the corresponding RU weights. | Resource | RU Weight | |:----------------|:-----------------| -| CPU | (1/3) RU/millisecond | -| Read IO | (1/64) RU/KB | +| CPU | 1/3 RU per millisecond | +| Read IO | 1/64 RU per KB | | Write IO | 1 RU/KB | | Basic overhead of a read request | 0.25 RU | | Basic overhead of a write request | 1.5 RU | -Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` requests that read `r2` KB data, `w1` write requests that write `w2` KB data, and the number of non-witness TiKV nodes in the cluster is `n`, then the formula for the total RUs consumed by the resource group is as follows: +Based on the above table, assuming that the TiKV time consumed by a resource group is `c` milliseconds, `r1` times of requests read `r2` KB data, `w1` times of write requests write `w2` KB data, and the number of non-witness TiKV nodes in the cluster is `n`. Then, the formula for the total RUs consumed by the resource group is as follows: `c`\* 1/3 + (`r1` \* 0.25 + `r2` \* 1/64) + (1.5 \* `w1` + `w2` \* 1 \* `n`) @@ -57,14 +59,14 @@ Based on the above table, assuming that the TiKV time consumed by a resource gro The resource control feature introduces two new global variables. -* TiDB: you can use the system variable [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) to control whether to enable flow control for a resource group. -* TiKV: you can use the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) to control whether to use request scheduling based on resource group quotas. +* TiDB: you can use the [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) system variable to control whether to enable flow control for resource groups. +* TiKV: you can use the [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) parameter to control whether to use request scheduling based on resource group quotas. > **Note:** > -> The parameter `resource-control.enabled` is disabled by default. It does not support dynamic modification. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it for TiDB Cloud Dedicated Tier clusters. +> In TiDB Cloud, the value of the `resource-control.enabled` parameter is `false` by default and does not support dynamic modification. If you need to enable it for TiDB Cloud Dedicated Tier clusters, contact [TiDB Cloud Support](/tidb-cloud/tidb-cloud-support.md). @@ -79,7 +81,7 @@ The results of the combinations of these two parameters are shown in the followi To create, modify, or delete a resource group, you need to have the `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. -You can create a resource group in the cluster by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind the users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). +You can create a resource group in the cluster by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). For an existing resource group, you can modify the read and write quota of the resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md). The quota changes to the resource group take effect immediately. @@ -88,7 +90,7 @@ You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements > **Note:** > > - When you bind a user to a resource group by using `CREATE USER` or `ALTER USER`, it will not take effect for the user's existing sessions, but only for the user's new sessions. -> - If a user is not bound to a resource group or is bound to a `default` resource group, the user's requests are not subject to TiDB's flow control restrictions. The `default` resource group is currently not visible to the user and cannot be created or modified. +> - If a user is not bound to a resource group or is bound to a `default` resource group, the user's requests are not subject to the flow control restrictions of TiDB. The `default` resource group is currently not visible to the user and cannot be created or modified. ### Step 1. Enable the resource control feature @@ -100,7 +102,7 @@ SET GLOBAL tidb_enable_resource_control = 'ON'; -In TiKV, set the parameter [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) to `true`. +Set the TiKV [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) parameter to `true`. @@ -120,7 +122,7 @@ The following is an example of how to create a resource group and bind users to CREATE RESOURCE GROUP IF NOT EXISTS rg1 RU_PER_SEC = 500 BURSTABLE; ``` -2. Create a resource group `rg2`. The RU backfill rate is 500 RUs per second and does not allow applications in this resource group to overrun resources. +2. Create a resource group `rg2`. The RU backfill rate is 600 RUs per second and does not allow applications in this resource group to overrun resources. ```sql CREATE RESOURCE GROUP IF NOT EXISTS rg2 RU_PER_SEC = 600; @@ -136,31 +138,33 @@ The following is an example of how to create a resource group and bind users to ALTER USER usr2 RESOURCE GROUP rg2; ``` -After you complete the above operations of creating resource groups and binding users, the resource consumption of newly created sessions will be controlled by the specified quota. If the system load is relatively high and there is no spare capacity, the resource consumption rate of `usr2` will be strictly controlled not to exceed the quota. Because `usr1` bound by `rg1` is configured with `BURSTABLE`, the consumption rate of `usr1` is allowed to exceed the quota. +After you complete the above operations of creating resource groups and binding users, the resource consumption of newly created sessions will be controlled by the specified quota. If the system workload is relatively high and there is no spare capacity, the resource consumption rate of `usr2` will be strictly controlled not to exceed the quota. Because `usr1` is bound by `rg1` with `BURSTABLE` configured, the consumption rate of `usr1` is allowed to exceed the quota. If the resource group corresponds to a request with insufficient quota, the client's request will wait. If the wait time is too long, the request will report an error. -## Monitoring and charts +## Monitoring metrics and charts > **Note:** > -> This section is not applicable to TiDB Cloud. +> This section is only applicable to on-premises TiDB. Currently, TiDB Cloud does not provide resource control metrics. -TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** panel. The metrics are detailed in the **Resource Control** section of [TiDB Important Monitoring Metrics](/grafana-tidb-dashboard.md). +TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** dashboard. The metrics are detailed in the **Resource Control** section of [TiDB Important Monitoring Metrics](/grafana-tidb-dashboard.md). + +TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). ## Tool Compatibility -The resource control feature is still in its experimental stage and does not impact the regular usage of data import, export, and other replication tools. BR, TiDB Lightning, and TiCDC do not currently support resource control related DDL processing, and their resource consumption is not limited by resource control. +The resource control feature is still in its experimental stage and does not impact the regular usage of data import, export, and other replication tools. BR, TiDB Lightning, and TiCDC do not currently support processing DDL operations related to resource control, and their resource consumption is not limited by resource control. ## Limitations Currently, the resource control feature has the following limitations: -* This feature only supports restriction and scheduling of read and write requests initiated by frontend clients. It does not support restriction and scheduling of background tasks such as `DDL` and `Auto Analyze`. +* This feature only supports flow control and scheduling of read and write requests initiated by foreground clients. It does not support flow control and scheduling of background tasks such as DDL operations and auto analyze. * Resource control incurs additional scheduling overhead. Therefore, there might be a slight performance degradation when this feature is enabled. ## See also diff --git a/tikv-configuration-file.md b/tikv-configuration-file.md index 577e279f1dc04..9293af45e63eb 100644 --- a/tikv-configuration-file.md +++ b/tikv-configuration-file.md @@ -2156,10 +2156,10 @@ To reduce write latency, TiKV periodically fetches and caches a batch of timesta ## resource-control -Configuration items related to resource control related to the TiKV storage layer. +Configuration items related to resource control of the TiKV storage layer. ### `enabled` New in v6.6.0 -+ Whether to support scheduling for user foreground read/write requests according to Request Unit (RU) of the corresponding resource group. For information about TiDB resource groups and resource control, see [TiDB resource control](/tidb-resource-control.md). -+ It is only meaningful to enable this configuration item if [`tidb_enable_resource_control](/system-variables.md#tidb_enable_resource_control-new-in-v660) is enabled on TiDB. When it is enabled, TiKV will use the priority queue to schedule the queued requests for user foreground read/write requests. The priority of scheduling is inversely related to the amount of resources already consumed by the resource group where the request is located, and positively related to the quota of the corresponding resource group. ++ Controls whether to enable scheduling for user foreground read/write requests according to [Request Unit (RU)](/tidb-resource-control.md#what-is-request-unit-ru) of the corresponding resource groups. For information about TiDB resource groups and resource control, see [TiDB resource control](/tidb-resource-control.md). ++ Enabling this configuration item only works when [`tidb_enable_resource_control](/system-variables.md#tidb_enable_resource_control-new-in-v660) is enabled on TiDB. When this configuration item is enabled, TiKV will use the priority queue to schedule the queued read/write requests from foreground users. The scheduling priority of a request is inversely related to the amount of resources already consumed by the resource group that receives this request, and positively related to the quota of the corresponding resource group. + Default value: `false`, which means to disable scheduling based on the RU of the resource group. \ No newline at end of file From 95a025d27b371bbce62a544114aa1ea434800b73 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Sat, 11 Feb 2023 22:36:59 +0800 Subject: [PATCH 39/63] Update grafana-resource-control-dashboard.md Co-authored-by: Grace Cai --- grafana-resource-control-dashboard.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index 1413e64cf71b4..a2dd9d8b05c72 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -24,7 +24,7 @@ This document describes some key monitoring metrics displayed on the Resource Co ## Metrics about resources -- KV Request Count: the number of KV requests is calculated in real time for each resource group. They are divided into read and write types. `total` is the sum of the KV requests involved by all Resource Groups. +- KV Request Count: the number of KV requests for each resource group, calculated per second. The requests are divided into read and write types. `total` is the sum of the KV requests for all Resource Groups. - KV Request Count Per Query: the average number of read and write KV requests by each SQL statement per second. It is obtained by dividing the above KV Request Count metric by the number of SQL statements executed per second. - Bytes Read: the amount of data read by each Resource Group, calculated per second. `total` is the sum of the data read by all Resource Groups. - Bytes Read Per Query: the average amount of data read by each SQL statement per second. It is obtained by dividing the above Bytes Read metric by the number of SQL statements executed per second. From 67efd7073f4f213e217897cdb80270e4fca26030 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Sat, 11 Feb 2023 22:37:39 +0800 Subject: [PATCH 40/63] Apply suggestions from code review --- tidb-resource-control.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index dc709a535f14e..08d0dd0f56f27 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -16,8 +16,6 @@ The TiDB resource control feature provides two layers of resource management cap - TiDB flow control: TiDB flow control uses the [token bucket algorithm](https://en.wikipedia.org/wiki/Token_bucket). If there are not enough tokens in a bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry might fail due to timeout. - TiKV scheduling: if [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to determine the priority of the read and write requests for each resource group. Based on the priorities, the storage layer uses the priority queue to schedule and process requests. -The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. - > **Note:** @@ -30,6 +28,8 @@ The introduction of the resource control feature is a milestone for TiDB. It can The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. +The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. + With this feature, you can: - Combine multiple small and medium-sized applications from different systems into a single TiDB cluster. When the workload of an application grows larger, it does not affect the normal operation of other applications. When the system workload is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. @@ -156,6 +156,8 @@ TiDB regularly collects runtime information about resource control and provides TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). +TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). + ## Tool Compatibility The resource control feature is still in its experimental stage and does not impact the regular usage of data import, export, and other replication tools. BR, TiDB Lightning, and TiCDC do not currently support processing DDL operations related to resource control, and their resource consumption is not limited by resource control. From 53e52b05ae89f8fbfe33441094a58d6275fd60a9 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Sat, 11 Feb 2023 22:38:16 +0800 Subject: [PATCH 41/63] Update grafana-tikv-dashboard.md --- grafana-tikv-dashboard.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/grafana-tikv-dashboard.md b/grafana-tikv-dashboard.md index 131e5a32810fb..07f0083175600 100644 --- a/grafana-tikv-dashboard.md +++ b/grafana-tikv-dashboard.md @@ -68,6 +68,9 @@ This section provides a detailed description of these key metrics on the **TiKV- - Average gRPC message duration: The average execution time of gRPC messages - gRPC batch size: The batch size of gRPC messages between TiDB and TiKV - Raft message batch size: The batch size of Raft messages between TiKV instances +- gRPC resource group duration: The QPS of gRPC request sources +- gRPC request sources duration: The execution time of gRPC request sources +- gRPC resource group QPS: The QPS of gRPC resource groups ### Thread CPU From 587cf9486a5aab0b432a108a4c730bd7b6700521 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Sat, 11 Feb 2023 23:01:43 +0800 Subject: [PATCH 42/63] Update information-schema-processlist.md --- information-schema/information-schema-processlist.md | 1 + 1 file changed, 1 insertion(+) diff --git a/information-schema/information-schema-processlist.md b/information-schema/information-schema-processlist.md index 05724f5946b66..919ac63b9da20 100644 --- a/information-schema/information-schema-processlist.md +++ b/information-schema/information-schema-processlist.md @@ -48,6 +48,7 @@ DESC processlist; ```sql SELECT * FROM processlist\G ``` + ```sql *************************** 1. row *************************** ID: 2300033189772525975 From 5532457c7bdd80cce1f90b2bac22e116c2cc8de1 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Sun, 12 Feb 2023 10:22:46 +0800 Subject: [PATCH 43/63] Update sql-statement-drop-resource-group.md --- sql-statements/sql-statement-drop-resource-group.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sql-statements/sql-statement-drop-resource-group.md b/sql-statements/sql-statement-drop-resource-group.md index 1ca96cf098efe..cf0d6c6acfac4 100644 --- a/sql-statements/sql-statement-drop-resource-group.md +++ b/sql-statements/sql-statement-drop-resource-group.md @@ -41,10 +41,10 @@ mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; +------+------------+-----------+ 1 row in set (0.01 sec) -mysql> DROP RESOURCE GROUP IF EXISTS rg1 ; +mysql> DROP RESOURCE GROUP IF EXISTS rg1; Query OK, 1 rows affected (0.09 sec) -mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; +mysql> SELECT * FROM information_schema.resource_groups WHERE NAME ='rg1'; Empty set (0.00 sec) ``` From 36e9fc7f530fc13eabb6ce920a22ce8537391115 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Wed, 15 Feb 2023 16:01:10 +0800 Subject: [PATCH 44/63] add the note that this feature does not support serverless tier --- .../information-schema-resource-groups.md | 8 ++++++++ .../sql-statement-alter-resource-group.md | 8 ++++++++ .../sql-statement-create-resource-group.md | 8 ++++++++ .../sql-statement-drop-resource-group.md | 8 ++++++++ .../sql-statement-show-create-resource-group.md | 8 ++++++++ tidb-resource-control.md | 14 +++++++------- 6 files changed, 47 insertions(+), 7 deletions(-) diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index a2993c385a9d1..172b567281999 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -9,6 +9,14 @@ summary: Learn the `RESOURCE_GROUPS` information_schema table. > > This feature is experimental and its form and usage might change in subsequent versions. + + +> **Note:** +> +> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). + + + The `RESOURCE_GROUPS` table shows the information about all resource groups. For more information, see [Use Resource Control to Achieve Resource Isolation](/tidb-resource-control.md). ```sql diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 56df867d26a66..6a13caa1b058f 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -5,6 +5,14 @@ summary: Learn the usage of ALTER RESOURCE GROUP in TiDB. # ALTER RESOURCE GROUP + + +> **Note:** +> +> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). + + + The `ALTER RESOURCE GROUP` statement is used to modify a resource group in a database. ## Synopsis diff --git a/sql-statements/sql-statement-create-resource-group.md b/sql-statements/sql-statement-create-resource-group.md index c721ff2fd5009..39ab52ac72e16 100644 --- a/sql-statements/sql-statement-create-resource-group.md +++ b/sql-statements/sql-statement-create-resource-group.md @@ -5,6 +5,14 @@ summary: Learn the usage of CREATE RESOURCE GROUP in TiDB. # CREATE RESOURCE GROUP + + +> **Note:** +> +> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). + + + You can use the `CREATE RESOURCE GROUP` statement to create a resource group. ## Synopsis diff --git a/sql-statements/sql-statement-drop-resource-group.md b/sql-statements/sql-statement-drop-resource-group.md index cf0d6c6acfac4..2edd398c59be8 100644 --- a/sql-statements/sql-statement-drop-resource-group.md +++ b/sql-statements/sql-statement-drop-resource-group.md @@ -5,6 +5,14 @@ summary: Learn the usage of DROP RESOURCE GROUP in TiDB. # DROP RESOURCE GROUP + + +> **Note:** +> +> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). + + + You can use the `DROP RESOURCE GROUP` statement to drop a resource group. ## Synopsis diff --git a/sql-statements/sql-statement-show-create-resource-group.md b/sql-statements/sql-statement-show-create-resource-group.md index 9b3cc33927338..ab3523a14df24 100644 --- a/sql-statements/sql-statement-show-create-resource-group.md +++ b/sql-statements/sql-statement-show-create-resource-group.md @@ -5,6 +5,14 @@ summary: Learn the usage of SHOW CREATE RESOURCE GROUP in TiDB. # SHOW CREATE RESOURCE GROUP + + +> **Note:** +> +> This feature is not available on [Serverless Tier clusters](/tidb-cloud/select-cluster-tier.md#serverless-tier-beta). + + + You can use the `SHOW CREATE RESOURCE GROUP` statement to view the current definition of a resource group. ## Synopsis diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 08d0dd0f56f27..215aa823d6ec1 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -9,13 +9,6 @@ summary: Learn how to use the resource control feature to control and schedule a > > This feature is experimental and its form and usage might change in subsequent versions. -As a cluster administrator, you can use the resource control feature to create resource groups, set read and write quotas for resource groups, and bind users to those groups. This allows the TiDB layer to control the flow of user read and write requests based on the quotas set for the resource groups, and allows the TiKV layer to schedule the requests based on the priority mapped to the read and write quota. By doing this, you can ensure resource isolation for your applications and meet quality of service (QoS) requirements. - -The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. - -- TiDB flow control: TiDB flow control uses the [token bucket algorithm](https://en.wikipedia.org/wiki/Token_bucket). If there are not enough tokens in a bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry might fail due to timeout. -- TiKV scheduling: if [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to determine the priority of the read and write requests for each resource group. Based on the priorities, the storage layer uses the priority queue to schedule and process requests. - > **Note:** @@ -24,6 +17,13 @@ The TiDB resource control feature provides two layers of resource management cap +As a cluster administrator, you can use the resource control feature to create resource groups, set read and write quotas for resource groups, and bind users to those groups. This allows the TiDB layer to control the flow of user read and write requests based on the quotas set for the resource groups, and allows the TiKV layer to schedule the requests based on the priority mapped to the read and write quota. By doing this, you can ensure resource isolation for your applications and meet quality of service (QoS) requirements. + +The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. + +- TiDB flow control: TiDB flow control uses the [token bucket algorithm](https://en.wikipedia.org/wiki/Token_bucket). If there are not enough tokens in a bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry might fail due to timeout. +- TiKV scheduling: if [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to determine the priority of the read and write requests for each resource group. Based on the priorities, the storage layer uses the priority queue to schedule and process requests. + ## Scenarios for resource control The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. From 58083ac7e1d7d187d2a986858509507c92cb9bc4 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Wed, 15 Feb 2023 21:53:06 +0800 Subject: [PATCH 45/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 215aa823d6ec1..029dc5c141c5e 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -77,6 +77,8 @@ The results of the combinations of these two parameters are shown in the followi | `resource-control.enabled`= true | Flow control and scheduling (recommended) | Invalid combination | | `resource-control.enabled`= false | Only flow control (not recommended) | The feature is disabled. | +For more information about the resource control mechanism, see [RFC: Global Resource Control in TiDB](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/edit?n= RFC_Global_Resource_Control_in_TiDB.docx#heading=h.sjp9cmjfszlu). + ## How to use resource control To create, modify, or delete a resource group, you need to have the `SUPER` or `RESOURCE_GROUP_ADMIN` privilege. From b25be98c2c28b92e1323fc0064270035e33db154 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 15 Feb 2023 22:03:56 +0800 Subject: [PATCH 46/63] Apply suggestions from code review Co-authored-by: Grace Cai --- grafana-resource-control-dashboard.md | 2 +- grafana-tikv-dashboard.md | 4 ++-- tidb-resource-control.md | 2 -- 3 files changed, 3 insertions(+), 5 deletions(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index a2dd9d8b05c72..e1437e348450c 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -19,7 +19,7 @@ This document describes some key monitoring metrics displayed on the Resource Co - RU Per Query: the average number of Request Units consumed by each SQL statement per second. It is obtained by dividing the above RU metric by the number of SQL statements executed per second. - RRU: the Read Request Unit consumption information of each resource group, calculated in real time. `total` is the sum of the Read Request Units consumed by all Resource Groups. - RRU Per Query: the average number of Read Request Units consumed by each SQL statement per second. It is obtained by dividing the above RRU metric by the number of SQL statements executed per second. -- WRU: the Write Request Unit consumption information of each resource group, calculated in real time. `total` is the sum of the Write Request Units consumed by all Resource Groups. +- WRU: the Write Request Unit consumption information of each resource group, calculated in real time. `total` is the sum of the Write Request Units consumed by all Resource Groups. - WRU Per Query: the average number of Write Request Units consumed by each SQL statement per second. It is obtained by dividing the above WRU metric by the number of SQL statements executed per second. ## Metrics about resources diff --git a/grafana-tikv-dashboard.md b/grafana-tikv-dashboard.md index 07f0083175600..69e92054bab8c 100644 --- a/grafana-tikv-dashboard.md +++ b/grafana-tikv-dashboard.md @@ -68,9 +68,9 @@ This section provides a detailed description of these key metrics on the **TiKV- - Average gRPC message duration: The average execution time of gRPC messages - gRPC batch size: The batch size of gRPC messages between TiDB and TiKV - Raft message batch size: The batch size of Raft messages between TiKV instances -- gRPC resource group duration: The QPS of gRPC request sources +- gRPC request sources QPS: The QPS of gRPC request sources - gRPC request sources duration: The execution time of gRPC request sources -- gRPC resource group QPS: The QPS of gRPC resource groups +- gRPC resource group QPS: The QPS of gRPC request sources by resource groups ### Thread CPU diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 029dc5c141c5e..d5c7973cff8e0 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -158,8 +158,6 @@ TiDB regularly collects runtime information about resource control and provides TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). -TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). - ## Tool Compatibility The resource control feature is still in its experimental stage and does not impact the regular usage of data import, export, and other replication tools. BR, TiDB Lightning, and TiCDC do not currently support processing DDL operations related to resource control, and their resource consumption is not limited by resource control. From 12ae27b2e79d6ec26c9c11abc30570f1d26c6675 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 15 Feb 2023 22:24:39 +0800 Subject: [PATCH 47/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 -- 1 file changed, 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index d5c7973cff8e0..197cce81d47c6 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -28,8 +28,6 @@ The TiDB resource control feature provides two layers of resource management cap The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. -The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. - With this feature, you can: - Combine multiple small and medium-sized applications from different systems into a single TiDB cluster. When the workload of an application grows larger, it does not affect the normal operation of other applications. When the system workload is low, busy applications can still be allocated the required system resources even if they exceed the set read and write quotas, so as to achieve the maximum utilization of resources. From 117ce01a2c10d28f93b3bd435898f169b2e3c056 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 15 Feb 2023 22:27:11 +0800 Subject: [PATCH 48/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 197cce81d47c6..3a3ad72586252 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -58,7 +58,7 @@ Based on the above table, assuming that the TiKV time consumed by a resource gro The resource control feature introduces two new global variables. * TiDB: you can use the [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) system variable to control whether to enable flow control for resource groups. -* TiKV: you can use the [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) parameter to control whether to use request scheduling based on resource group quotas. +* TiKV: you can use the [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) parameter to control whether to use request scheduling based on resource groups. From 661d1eb8719e453e05acfd65cb14fca6c710d1d5 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Wed, 15 Feb 2023 22:35:04 +0800 Subject: [PATCH 49/63] Apply suggestions from code review --- tidb-resource-control.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 3a3ad72586252..a9d754cfab8fb 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -64,7 +64,7 @@ The resource control feature introduces two new global variables. > **Note:** > -> In TiDB Cloud, the value of the `resource-control.enabled` parameter is `false` by default and does not support dynamic modification. If you need to enable it for TiDB Cloud Dedicated Tier clusters, contact [TiDB Cloud Support](/tidb-cloud/tidb-cloud-support.md). +> In TiDB Cloud, the value of the `resource-control.enabled` parameter is `false` by default and does not support dynamic modification. If you need to enable it for TiDB Cloud Dedicated Tier clusters, contact [TiDB Cloud Support](/tidb-cloud/tidb-cloud-support.md). @@ -83,7 +83,7 @@ To create, modify, or delete a resource group, you need to have the `SUPER` or ` You can create a resource group in the cluster by using [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md), and then bind users to a specific resource group by using [`CREATE USER`](/sql-statements/ sql-statement-create-user.md) or [`ALTER USER`](/sql-statements/sql-statement-alter-user.md). -For an existing resource group, you can modify the read and write quota of the resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md). The quota changes to the resource group take effect immediately. +For an existing resource group, you can modify the `RU_PER_SEC` option (the rate of RU backfilling per second) of the resource group by using [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md). The changes to the resource group take effect immediately. You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md). From ccbd572daa738a85a2e579cb4b8c0a24904604ff Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Thu, 16 Feb 2023 10:10:27 +0800 Subject: [PATCH 50/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index a9d754cfab8fb..83948d1118d2e 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -75,7 +75,7 @@ The results of the combinations of these two parameters are shown in the followi | `resource-control.enabled`= true | Flow control and scheduling (recommended) | Invalid combination | | `resource-control.enabled`= false | Only flow control (not recommended) | The feature is disabled. | -For more information about the resource control mechanism, see [RFC: Global Resource Control in TiDB](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/edit?n= RFC_Global_Resource_Control_in_TiDB.docx#heading=h.sjp9cmjfszlu). +For more information about the resource control mechanism and parameters, see [RFC: Global Resource Control in TiDB](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/edit?n= RFC_Global_Resource_Control_in_TiDB.docx#heading=h.sjp9cmjfszlu). ## How to use resource control From 3b69ca4653344c0b282f5b9e6e49f3f6e676d55c Mon Sep 17 00:00:00 2001 From: BornChanger <97348524+BornChanger@users.noreply.github.com> Date: Thu, 16 Feb 2023 16:52:32 +0800 Subject: [PATCH 51/63] Update information-schema/information-schema-resource-groups.md --- information-schema/information-schema-resource-groups.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index 172b567281999..260c65b653037 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -32,7 +32,7 @@ DESC resource_groups; | RU_PER_SEC | bigint(21) | YES | | NULL | | | BURSTABLE | varchar(3) | YES | | NULL | | +------------+-------------+------+------+---------+-------+ -4 rows in set (0.00 sec) +3 rows in set (0.00 sec) ``` ## Examples From ff329dd776eda4eac502479a0ce1f931751a5bc7 Mon Sep 17 00:00:00 2001 From: BornChanger <97348524+BornChanger@users.noreply.github.com> Date: Thu, 16 Feb 2023 16:53:14 +0800 Subject: [PATCH 52/63] Update information-schema/information-schema-resource-groups.md --- information-schema/information-schema-resource-groups.md | 1 + 1 file changed, 1 insertion(+) diff --git a/information-schema/information-schema-resource-groups.md b/information-schema/information-schema-resource-groups.md index 260c65b653037..8b094de03661e 100644 --- a/information-schema/information-schema-resource-groups.md +++ b/information-schema/information-schema-resource-groups.md @@ -53,6 +53,7 @@ mysql> SELECT * FROM information_schema.resource_groups WHERE NAME = 'rg1'; +------+------------+-----------+ | rg1 | 1000 | NO | +------+------------+-----------+ +1 row in set (0.00 sec) ``` The descriptions of the columns in the `RESOURCE_GROUPS` table are as follows: From 8325f7ff1f6f7995d4088aa6738c79cf342488dc Mon Sep 17 00:00:00 2001 From: BornChanger <97348524+BornChanger@users.noreply.github.com> Date: Thu, 16 Feb 2023 16:58:30 +0800 Subject: [PATCH 53/63] Update sql-statements/sql-statement-alter-resource-group.md --- sql-statements/sql-statement-alter-resource-group.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 6a13caa1b058f..406691259e5cf 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -19,7 +19,7 @@ The `ALTER RESOURCE GROUP` statement is used to modify a resource group in a dat ```ebnf+diagram AlterResourceGroupStmt: - "ALTER" "RESOURCE" "GROUP" IfNotExists ResourceGroupName ResourceGroupOptionList + "ALTER" "RESOURCE" "GROUP" IfExists ResourceGroupName ResourceGroupOptionList IfNotExists ::= ('IF' 'EXISTS')? From 7b38d6807ec24ce3c8ae15af79bd6f7de26f7290 Mon Sep 17 00:00:00 2001 From: BornChanger <97348524+BornChanger@users.noreply.github.com> Date: Thu, 16 Feb 2023 16:58:42 +0800 Subject: [PATCH 54/63] Update sql-statements/sql-statement-alter-resource-group.md --- sql-statements/sql-statement-alter-resource-group.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sql-statements/sql-statement-alter-resource-group.md b/sql-statements/sql-statement-alter-resource-group.md index 406691259e5cf..088f2b5759abe 100644 --- a/sql-statements/sql-statement-alter-resource-group.md +++ b/sql-statements/sql-statement-alter-resource-group.md @@ -21,7 +21,7 @@ The `ALTER RESOURCE GROUP` statement is used to modify a resource group in a dat AlterResourceGroupStmt: "ALTER" "RESOURCE" "GROUP" IfExists ResourceGroupName ResourceGroupOptionList -IfNotExists ::= +IfExists ::= ('IF' 'EXISTS')? ResourceGroupName: From 3a38be124f0cf1926b2ce8108a99651db266365a Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Fri, 17 Feb 2023 09:07:41 +0800 Subject: [PATCH 55/63] Apply suggestions from code review Co-authored-by: Grace Cai --- tidb-resource-control.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 83948d1118d2e..c07d055ea03db 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -140,7 +140,7 @@ The following is an example of how to create a resource group and bind users to After you complete the above operations of creating resource groups and binding users, the resource consumption of newly created sessions will be controlled by the specified quota. If the system workload is relatively high and there is no spare capacity, the resource consumption rate of `usr2` will be strictly controlled not to exceed the quota. Because `usr1` is bound by `rg1` with `BURSTABLE` configured, the consumption rate of `usr1` is allowed to exceed the quota. -If the resource group corresponds to a request with insufficient quota, the client's request will wait. If the wait time is too long, the request will report an error. +If the resource group corresponds to a request with an insufficient quota, the client's request will wait. If the wait time is too long, the request will report an error. ## Monitoring metrics and charts @@ -156,7 +156,7 @@ TiDB regularly collects runtime information about resource control and provides TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). -## Tool Compatibility +## Tool compatibility The resource control feature is still in its experimental stage and does not impact the regular usage of data import, export, and other replication tools. BR, TiDB Lightning, and TiCDC do not currently support processing DDL operations related to resource control, and their resource consumption is not limited by resource control. From 56086f06dcd95c952a5b773c5b4deb64a1e8fb55 Mon Sep 17 00:00:00 2001 From: xixirangrang Date: Fri, 17 Feb 2023 11:02:40 +0800 Subject: [PATCH 56/63] Apply suggestions from code review --- tidb-resource-control.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index c07d055ea03db..1843fbe16c67c 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -90,7 +90,7 @@ You can delete a resource group by using [`DROP RESOURCE GROUP`](/sql-statements > **Note:** > > - When you bind a user to a resource group by using `CREATE USER` or `ALTER USER`, it will not take effect for the user's existing sessions, but only for the user's new sessions. -> - If a user is not bound to a resource group or is bound to a `default` resource group, the user's requests are not subject to the flow control restrictions of TiDB. The `default` resource group is currently not visible to the user and cannot be created or modified. +> - If a user is not bound to a resource group or is bound to a `default` resource group, the user's requests are not subject to the flow control restrictions of TiDB. The `default` resource group is currently not visible to the user and cannot be created or modified. You cannot view it with `SHOW CREATE RESOURCE GROUP` or `SELECT * FROM information_schema.resource_groups`. But you can view it through the `mysql.user` table. ### Step 1. Enable the resource control feature @@ -140,7 +140,7 @@ The following is an example of how to create a resource group and bind users to After you complete the above operations of creating resource groups and binding users, the resource consumption of newly created sessions will be controlled by the specified quota. If the system workload is relatively high and there is no spare capacity, the resource consumption rate of `usr2` will be strictly controlled not to exceed the quota. Because `usr1` is bound by `rg1` with `BURSTABLE` configured, the consumption rate of `usr1` is allowed to exceed the quota. -If the resource group corresponds to a request with an insufficient quota, the client's request will wait. If the wait time is too long, the request will report an error. +If there are too many requests that result in insufficient resources for the resource group, the client's requests will wait. If the wait time is too long, the requests will report an error. ## Monitoring metrics and charts From 6ae50af8ccdba51e6f0104ed51c98ac2c1805070 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 17 Feb 2023 11:15:55 +0800 Subject: [PATCH 57/63] Update TOC-tidb-cloud.md --- TOC-tidb-cloud.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/TOC-tidb-cloud.md b/TOC-tidb-cloud.md index dda16cb26a4e9..4df944f568cab 100644 --- a/TOC-tidb-cloud.md +++ b/TOC-tidb-cloud.md @@ -98,6 +98,7 @@ - [Scale a TiDB Cluster](/tidb-cloud/scale-tidb-cluster.md) - [Pause or Resume a TiDB Cluster](/tidb-cloud/pause-or-resume-tidb-cluster.md) - [Upgrade a TiDB Cluster](/tidb-cloud/upgrade-tidb-cluster.md) + - [Resource Control](/tidb-resource-control.md) - [Delete a TiDB Cluster](/tidb-cloud/delete-tidb-cluster.md) - Migrate or Import Data - [Overview](/tidb-cloud/tidb-cloud-migration-overview.md) @@ -254,6 +255,7 @@ - [`ADMIN SHOW DDL [JOBS|JOB QUERIES]`](/sql-statements/sql-statement-admin-show-ddl.md) - [`ALTER DATABASE`](/sql-statements/sql-statement-alter-database.md) - [`ALTER INDEX`](/sql-statements/sql-statement-alter-index.md) + - [`ALTER RESOURCE GROUP`](/sql-statements/sql-statement-alter-resource-group.md) - [`ALTER TABLE`](/sql-statements/sql-statement-alter-table.md) - [`ALTER TABLE COMPACT`](/sql-statements/sql-statement-alter-table-compact.md) - [`ALTER USER`](/sql-statements/sql-statement-alter-user.md) @@ -267,6 +269,7 @@ - [`CREATE [GLOBAL|SESSION] BINDING`](/sql-statements/sql-statement-create-binding.md) - [`CREATE DATABASE`](/sql-statements/sql-statement-create-database.md) - [`CREATE INDEX`](/sql-statements/sql-statement-create-index.md) + - [`CREATE RESOURCE GROUP`](/sql-statements/sql-statement-create-resource-group.md) - [`CREATE ROLE`](/sql-statements/sql-statement-create-role.md) - [`CREATE SEQUENCE`](/sql-statements/sql-statement-create-sequence.md) - [`CREATE TABLE LIKE`](/sql-statements/sql-statement-create-table-like.md) @@ -282,6 +285,7 @@ - [`DROP COLUMN`](/sql-statements/sql-statement-drop-column.md) - [`DROP DATABASE`](/sql-statements/sql-statement-drop-database.md) - [`DROP INDEX`](/sql-statements/sql-statement-drop-index.md) + - [`DROP RESOURCE GROUP`](/sql-statements/sql-statement-drop-resource-group.md) - [`DROP ROLE`](/sql-statements/sql-statement-drop-role.md) - [`DROP SEQUENCE`](/sql-statements/sql-statement-drop-sequence.md) - [`DROP STATS`](/sql-statements/sql-statement-drop-stats.md) @@ -327,6 +331,7 @@ - [`SHOW COLLATION`](/sql-statements/sql-statement-show-collation.md) - [`SHOW [FULL] COLUMNS FROM`](/sql-statements/sql-statement-show-columns-from.md) - [`SHOW CREATE DATABASE`](/sql-statements/sql-statement-show-create-database.md) + - [`SHOW CREATE RESOURCE GROUP`](/sql-statements/sql-statement-show-create-resource-group.md) - [`SHOW CREATE SEQUENCE`](/sql-statements/sql-statement-show-create-sequence.md) - [`SHOW CREATE TABLE`](/sql-statements/sql-statement-show-create-table.md) - [`SHOW CREATE USER`](/sql-statements/sql-statement-show-create-user.md) @@ -444,6 +449,7 @@ - [`PARTITIONS`](/information-schema/information-schema-partitions.md) - [`PROCESSLIST`](/information-schema/information-schema-processlist.md) - [`REFERENTIAL_CONSTRAINTS`](/information-schema/information-schema-referential-constraints.md) + - [`RESOURCE_GROUPS`](/information-schema/information-schema-resource-groups.md) - [`SCHEMATA`](/information-schema/information-schema-schemata.md) - [`SEQUENCES`](/information-schema/information-schema-sequences.md) - [`SESSION_VARIABLES`](/information-schema/information-schema-session-variables.md) From e17d38b666462e92f5e3f01ad846db6f52280e60 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 17 Feb 2023 11:33:53 +0800 Subject: [PATCH 58/63] Update TOC-tidb-cloud.md --- TOC-tidb-cloud.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/TOC-tidb-cloud.md b/TOC-tidb-cloud.md index 4df944f568cab..522afa658bc43 100644 --- a/TOC-tidb-cloud.md +++ b/TOC-tidb-cloud.md @@ -98,7 +98,6 @@ - [Scale a TiDB Cluster](/tidb-cloud/scale-tidb-cluster.md) - [Pause or Resume a TiDB Cluster](/tidb-cloud/pause-or-resume-tidb-cluster.md) - [Upgrade a TiDB Cluster](/tidb-cloud/upgrade-tidb-cluster.md) - - [Resource Control](/tidb-resource-control.md) - [Delete a TiDB Cluster](/tidb-cloud/delete-tidb-cluster.md) - Migrate or Import Data - [Overview](/tidb-cloud/tidb-cloud-migration-overview.md) @@ -483,6 +482,7 @@ - [Table Filter](/table-filter.md) - [Troubleshoot Inconsistency Between Data and Indexes](/troubleshoot-data-inconsistency-errors.md) - [Serverless Tier Limitations](/tidb-cloud/serverless-tier-limitations.md) + - [Resource Control](/tidb-resource-control.md) - FAQs - [TiDB Cloud FAQs](/tidb-cloud/tidb-cloud-faq.md) - [Serverless Tier FAQs](/tidb-cloud/serverless-tier-faqs.md) From 81632ad5d2346d0da1ec0434a4662b48d03771e9 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 17 Feb 2023 11:39:14 +0800 Subject: [PATCH 59/63] Update tidb-resource-control.md --- tidb-resource-control.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 1843fbe16c67c..a6ea11ff5f940 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -75,7 +75,7 @@ The results of the combinations of these two parameters are shown in the followi | `resource-control.enabled`= true | Flow control and scheduling (recommended) | Invalid combination | | `resource-control.enabled`= false | Only flow control (not recommended) | The feature is disabled. | -For more information about the resource control mechanism and parameters, see [RFC: Global Resource Control in TiDB](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/edit?n= RFC_Global_Resource_Control_in_TiDB.docx#heading=h.sjp9cmjfszlu). +For more information about the resource control mechanism and parameters, see [RFC: Global Resource Control in TiDB](https://docs.google.com/document/d/1sV5EVv8Cdpc6aBCDihc2akpE0iuantPf/edit?n=RFC_Global_Resource_Control_in_TiDB.docx#heading=h.sjp9cmjfszlu). ## How to use resource control From f2e351bd7621436880a03af7803a73dff35ec8c6 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 17 Feb 2023 11:49:21 +0800 Subject: [PATCH 60/63] Update grafana-resource-control-dashboard.md --- grafana-resource-control-dashboard.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index e1437e348450c..fd4002f75078a 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -9,7 +9,7 @@ If you use TiUP to deploy the TiDB cluster, the monitoring system (Prometheus & The Grafana dashboard is divided into a series of sub dashboards which include Overview, PD, TiDB, TiKV, Node\_exporter, Disk Performance, and Performance\_overview. -If your cluster has used the [Resource Control](tidb-resource-control.md) feature, you can get an overview of the resource consumption status from the Resource Control dashboard. +If your cluster has used the [Resource Control](/tidb-resource-control.md) feature, you can get an overview of the resource consumption status from the Resource Control dashboard. This document describes some key monitoring metrics displayed on the Resource Control dashboard. From c2c0dbe330ef0a0b1c6c46b6477b174866518409 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 17 Feb 2023 11:54:14 +0800 Subject: [PATCH 61/63] Update grafana-resource-control-dashboard.md --- grafana-resource-control-dashboard.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/grafana-resource-control-dashboard.md b/grafana-resource-control-dashboard.md index fd4002f75078a..c75719342b2b3 100644 --- a/grafana-resource-control-dashboard.md +++ b/grafana-resource-control-dashboard.md @@ -9,7 +9,7 @@ If you use TiUP to deploy the TiDB cluster, the monitoring system (Prometheus & The Grafana dashboard is divided into a series of sub dashboards which include Overview, PD, TiDB, TiKV, Node\_exporter, Disk Performance, and Performance\_overview. -If your cluster has used the [Resource Control](/tidb-resource-control.md) feature, you can get an overview of the resource consumption status from the Resource Control dashboard. +If your cluster has used the [Resource Control](/tidb-resource-control.md) feature, you can get an overview of the resource consumption status from the Resource Control dashboard. This document describes some key monitoring metrics displayed on the Resource Control dashboard. From d989f7d7d2c009b4f5c1f344dca564bd1a46d48b Mon Sep 17 00:00:00 2001 From: Grace Cai Date: Fri, 17 Feb 2023 13:38:00 +0800 Subject: [PATCH 62/63] add the customcontent for TiDB Cloud --- tidb-resource-control.md | 36 +++++++++++++++++++++++++++++------- 1 file changed, 29 insertions(+), 7 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index a6ea11ff5f940..327977c58fc52 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -22,8 +22,19 @@ As a cluster administrator, you can use the resource control feature to create r The TiDB resource control feature provides two layers of resource management capabilities: the flow control capability at the TiDB layer and the priority scheduling capability at the TiKV layer. The two capabilities can be enabled separately or simultaneously. See the [Parameters for resource control](#parameters-for-resource-control) for details. - TiDB flow control: TiDB flow control uses the [token bucket algorithm](https://en.wikipedia.org/wiki/Token_bucket). If there are not enough tokens in a bucket, and the resource group does not specify the `BURSTABLE` option, the requests to the resource group will wait for the token bucket to backfill the tokens and retry. The retry might fail due to timeout. + + + - TiKV scheduling: if [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to determine the priority of the read and write requests for each resource group. Based on the priorities, the storage layer uses the priority queue to schedule and process requests. + + + + +- TiKV scheduling: for on-premises TiDB, if the `resource-control.enabled` parameter is enabled, TiKV uses the value of `RU_PER_SEC` of each resource group to determine the priority of the read and write requests for each resource group. Based on the priorities, the storage layer uses the priority queue to schedule and process requests. For TiDB Cloud, the value of the `resource-control.enabled` parameter is `false` by default and does not support dynamic modification. If you need to enable it for TiDB Cloud Dedicated Tier clusters, contact [TiDB Cloud Support](/tidb-cloud/tidb-cloud-support.md). + + + ## Scenarios for resource control The introduction of the resource control feature is a milestone for TiDB. It can divide a distributed database cluster into multiple logical units. Even if an individual unit overuses resources, it does not crowd out the resources needed by other units. @@ -58,13 +69,16 @@ Based on the above table, assuming that the TiKV time consumed by a resource gro The resource control feature introduces two new global variables. * TiDB: you can use the [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) system variable to control whether to enable flow control for resource groups. + + + * TiKV: you can use the [`resource-control.enabled`](/tikv-configuration-file.md#resource-control) parameter to control whether to use request scheduling based on resource groups. + + -> **Note:** -> -> In TiDB Cloud, the value of the `resource-control.enabled` parameter is `false` by default and does not support dynamic modification. If you need to enable it for TiDB Cloud Dedicated Tier clusters, contact [TiDB Cloud Support](/tidb-cloud/tidb-cloud-support.md). +* TiKV: For on-premises TiDB, you can use the `resource-control.enabled` parameter to control whether to use request scheduling based on resource group quotas. For TiDB Cloud, the value of the `resource-control.enabled` parameter is `false` by default and does not support dynamic modification. If you need to enable it for TiDB Cloud Dedicated Tier clusters, contact [TiDB Cloud Support](/tidb-cloud/tidb-cloud-support.md). @@ -108,7 +122,7 @@ Set the TiKV [`resource-control.enabled`](/tikv-configuration-file.md#resource-c -In TiKV, set the parameter [`resource-control.enabled`](https://docs.pingcap.com/tidb/stable/tikv-configuration-file#resource_control) to `true`. The parameter `resource-control.enabled` is disabled by default. You need to contact the [PingCAP support team](/tidb-cloud/tidb-cloud-support.md) to enable it. +For on-premises TiDB, set the TiKV `resource-control.enabled` parameter to `true`. For TiDB Cloud, the value of the `resource-control.enabled` parameter is `false` by default and does not support dynamic modification. If you need to enable it for TiDB Cloud Dedicated Tier clusters, contact [TiDB Cloud Support](/tidb-cloud/tidb-cloud-support.md). @@ -144,17 +158,25 @@ If there are too many requests that result in insufficient resources for the res ## Monitoring metrics and charts + + +TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** dashboard. The metrics are detailed in the **Resource Control** section of [TiDB Important Monitoring Metrics](/grafana-tidb-dashboard.md). + +TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). + + + > **Note:** > > This section is only applicable to on-premises TiDB. Currently, TiDB Cloud does not provide resource control metrics. - +TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** dashboard. -TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** dashboard. The metrics are detailed in the **Resource Control** section of [TiDB Important Monitoring Metrics](/grafana-tidb-dashboard.md). +TiKV also records the request QPS from different resource groups in Grafana's **TiKV** dashboard. -TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). + ## Tool compatibility From c23efb2921847d49ce6e747741e3a7874ecacb45 Mon Sep 17 00:00:00 2001 From: xixirangrang <35301108+hfxsd@users.noreply.github.com> Date: Fri, 17 Feb 2023 13:52:13 +0800 Subject: [PATCH 63/63] Update tidb-resource-control.md --- tidb-resource-control.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tidb-resource-control.md b/tidb-resource-control.md index 327977c58fc52..fb3af5fb7f6fa 100644 --- a/tidb-resource-control.md +++ b/tidb-resource-control.md @@ -68,7 +68,7 @@ Based on the above table, assuming that the TiKV time consumed by a resource gro The resource control feature introduces two new global variables. -* TiDB: you can use the [`tidb_enable_resource_control`](/system-variables.md#tidb-tidb_enable_resource_control-new-in-v660) system variable to control whether to enable flow control for resource groups. +* TiDB: you can use the [`tidb_enable_resource_control`](/system-variables.md#tidb_enable_resource_control-new-in-v660) system variable to control whether to enable flow control for resource groups. @@ -162,7 +162,7 @@ If there are too many requests that result in insufficient resources for the res TiDB regularly collects runtime information about resource control and provides visual charts of the metrics in Grafana's **TiDB** > **Resource Control** dashboard. The metrics are detailed in the **Resource Control** section of [TiDB Important Monitoring Metrics](/grafana-tidb-dashboard.md). -TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#gRPC). +TiKV also records the request QPS from different resource groups. For more details, see [TiKV Monitoring Metrics Detail](/grafana-tikv-dashboard.md#grpc).