Skip to content

Pick "[enhance](Cooldown) Use config to control whether use cooldown replica for scanning first" #38322

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
yiguolei merged 1 commit into from
Jul 25, 2024
Merged

Pick "[enhance](Cooldown) Use config to control whether use cooldown replica for scanning first" #38322

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions fe/fe-common/src/main/java/org/apache/doris/common/Config.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2735,6 +2735,9 @@ public static boolean isNotCloudMode() {
"Stream_Load When importing, the maximum length of label is limited"})
public static int label_regex_length = 128;

@ConfField(mutable = true)
public static boolean enable_cooldown_replica_affinity = true;

//==========================================================================
// end of cloud config
//==========================================================================
Expand Down
42 changes: 22 additions & 20 deletions fe/fe-core/src/main/java/org/apache/doris/planner/OlapScanNode.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -836,27 +836,29 @@ private void addScanRangeLocations(Partition partition,
}
}

final long coolDownReplicaId = tablet.getCooldownReplicaId();
// we prefer to query using cooldown replica to make sure the cache is fully utilized
// for example: consider there are 3BEs(A,B,C) and each has one replica for tablet X. and X
// is now under cooldown
// first time we choose BE A, and A will download data into cache while the other two's cache is empty
// second time we choose BE B, this time B will be cached, C is still empty
// third time we choose BE C, after this time all replica is cached
// but it means we will do 3 S3 IO to get the data which will bring 3 slow query
if (-1L != coolDownReplicaId) {
final Optional<Replica> replicaOptional = replicas.stream()
.filter(r -> r.getId() == coolDownReplicaId).findAny();
replicaOptional.ifPresent(
r -> {
Backend backend = Env.getCurrentSystemInfo()
.getBackend(r.getBackendId());
if (backend != null && backend.isAlive()) {
replicas.clear();
replicas.add(r);
if (Config.enable_cooldown_replica_affinity) {
final long coolDownReplicaId = tablet.getCooldownReplicaId();
// we prefer to query using cooldown replica to make sure the cache is fully utilized
// for example: consider there are 3BEs(A,B,C) and each has one replica for tablet X. and X
// is now under cooldown
// first time we choose BE A, and A will download data into cache while the other two's cache is empty
// second time we choose BE B, this time B will be cached, C is still empty
// third time we choose BE C, after this time all replica is cached
// but it means we will do 3 S3 IO to get the data which will bring 3 slow query
if (-1L != coolDownReplicaId) {
final Optional<Replica> replicaOptional = replicas.stream()
.filter(r -> r.getId() == coolDownReplicaId).findAny();
replicaOptional.ifPresent(
r -> {
Backend backend = Env.getCurrentSystemInfo()
.getBackend(r.getBackendId());
if (backend != null && backend.isAlive()) {
replicas.clear();
replicas.add(r);
}
}
}
);
);
}
}
boolean tabletIsNull = true;
boolean collectedStat = false;
Expand Down