[opt](resource-tag) root and admin user can use any resource tag by default

docs/en/docs/admin-manual/multi-tenant.md

@@ -138,7 +138,7 @@ Node resource division refers to setting tags for BE nodes in a Doris cluster, a
 
     After the setting is complete, when user1 initiates a query on the UserTable table, it will only access the data copy on the nodes in the `group_a` resource group, and the query will only use the node computing resources in the `group_a` resource group. The query of user3 can use copies and computing resources in any resource group.
 
-    > Note: By default, the user's `resource_tags.location` attribute is empty. In versions prior to 2.0.2 (inclusive), by default, users are not restricted by tags and can use any resource group. After version 2.0.3, users can only use the `default` resource group by default.
+    > Note: By default, the user's `resource_tags.location` attribute is empty. In versions prior to 2.0.2 (inclusive), by default, users are not restricted by tags and can use any resource group. After version 2.0.3, normal users can only use the `default` resource group by default. Root and Admin user can use any resource group.
 
     In this way, we have achieved physical resource isolation for different user queries by dividing nodes and restricting user resource usage. Furthermore, we can create different users for different business departments and restrict each user from using different resource groups. In order to avoid the use of resource interference between different business parts. For example, there is a business table in the cluster that needs to be shared by all 9 business departments, but it is hoped that resource preemption between different departments can be avoided as much as possible. Then we can create 3 copies of this table and store them in 3 resource groups. Next, we create 9 users for 9 business departments, and limit the use of one resource group for every 3 users. In this way, the degree of competition for resources is reduced from 9 to 3.
 
@@ -268,4 +268,4 @@ Through the above 4 steps, we can smoothly use the resource division function af
 
    The table creation statements for table2, table3, and table4 do not need to specify `replication_allocation` again.
 
-   Note: Changing the replica distribution policy of the database will not affect existing tables.
+   Note: Changing the replica distribution policy of the database will not affect existing tables.

docs/zh-CN/docs/admin-manual/multi-tenant.md

@@ -138,7 +138,7 @@ FE 不参与用户数据的处理计算等工作，因此是一个资源消耗
 
    设置完成后，user1 在发起对 UserTable 表的查询时，只会访问 `group_a` 资源组内节点上的数据副本，并且查询仅会使用 `group_a` 资源组内的节点计算资源。而 user3 的查询可以使用任意资源组内的副本和计算资源。
 
-   > 注：默认情况下，用户的 `resource_tags.location` 属性为空，在2.0.2（含）之前的版本中，默认情况下，用户不受 tag 的限制，可以使用任意资源组。在 2.0.3 版本之后，默认情况下，用户只能使用 `default` 资源组。
+   > 注：默认情况下，用户的 `resource_tags.location` 属性为空，在2.0.2（含）之前的版本中，默认情况下，用户不受 tag 的限制，可以使用任意资源组。在 2.0.3 版本之后，默认情况下，普通用户只能使用 `default` 资源组。root 和 admin 用户可以使用任意资源组。
 
    这样，我们通过对节点的划分，以及对用户的资源使用限制，实现了不同用户查询上的物理资源隔离。更进一步，我们可以给不同的业务部门创建不同的用户，并限制每个用户使用不同的资源组。以避免不同业务部分之间使用资源干扰。比如集群内有一张业务表需要共享给所有9个业务部门使用，但是希望能够尽量避免不同部门之间的资源抢占。则我们可以为这张表创建3个副本，分别存储在3个资源组中。接下来，我们为9个业务部门创建9个用户，每3个用户限制使用一个资源组。这样，资源的竞争程度就由9降低到了3。
 
@@ -266,4 +266,4 @@ FE 不参与用户数据的处理计算等工作，因此是一个资源消耗
 
    table2，table3,table4的建表语句无需再指定`replication_allocation`。
 
-   注意事项：更改database的副本分布策略不会对已有的table产生影响。
+   注意事项：更改database的副本分布策略不会对已有的table产生影响。

fe/fe-core/src/main/java/org/apache/doris/mysql/privilege/CommonUserProperties.java

@@ -110,10 +110,6 @@ public void setResourceTags(Set<Tag> resourceTags) {
     }
 
     public Set<Tag> getResourceTags() {
-        // If resource tags in user properties is empty, use default backend tag.
-        if (resourceTags.isEmpty()) {
-            return Sets.newHashSet(Tag.DEFAULT_BACKEND_TAG);
-        }
         return resourceTags;
     }
 

fe/fe-core/src/main/java/org/apache/doris/mysql/privilege/UserPropertyMgr.java

@@ -29,6 +29,7 @@
 import org.apache.doris.resource.Tag;
 
 import com.google.common.collect.Maps;
+import com.google.common.collect.Sets;
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
 
@@ -123,7 +124,15 @@ public Set<Tag> getResourceTags(String qualifiedUser) {
         if (existProperty == null) {
             return UserProperty.INVALID_RESOURCE_TAGS;
         }
-        return existProperty.getCopiedResourceTags();
+        Set<Tag> tags = existProperty.getCopiedResourceTags();
+        // only root and admin can return empty tag.
+        // empty tag means user can access all backends.
+        // for normal user, if tag is empty, use default tag.
+        if (tags.isEmpty() && !(qualifiedUser.equalsIgnoreCase(Auth.ROOT_USER)
+                || qualifiedUser.equalsIgnoreCase(Auth.ADMIN_USER))) {
+            tags = Sets.newHashSet(Tag.DEFAULT_BACKEND_TAG);
+        }
+        return tags;
     }
 
     public Pair<String, DppConfig> getLoadClusterInfo(String qualifiedUser, String cluster) throws DdlException {

fe/fe-core/src/test/java/org/apache/doris/catalog/UserPropertyTest.java

@@ -24,7 +24,6 @@
 import org.apache.doris.common.UserException;
 import org.apache.doris.load.DppConfig;
 import org.apache.doris.mysql.privilege.UserProperty;
-import org.apache.doris.resource.Tag;
 
 import com.google.common.collect.Lists;
 import com.google.common.collect.Sets;
@@ -118,7 +117,7 @@ public void testUpdate() throws UserException {
         Assert.assertEquals(new String[]{"rule1", "rule2"}, userProperty.getSqlBlockRules());
         Assert.assertEquals(2, userProperty.getCpuResourceLimit());
         Assert.assertEquals(500, userProperty.getQueryTimeout());
-        Assert.assertEquals(Sets.newHashSet(Tag.DEFAULT_BACKEND_TAG), userProperty.getCopiedResourceTags());
+        Assert.assertEquals(Sets.newHashSet(), userProperty.getCopiedResourceTags());
 
         // fetch property
         List<List<String>> rows = userProperty.fetchProperty();

fe/fe-core/src/test/java/org/apache/doris/planner/ResourceTagQueryTest.java

@@ -196,7 +196,7 @@ public void test() throws Exception {
         OlapTable tbl = (OlapTable) db.getTableNullable("tbl1");
 
         Set<Tag> userTags = Env.getCurrentEnv().getAuth().getResourceTags(Auth.ROOT_USER);
-        Assert.assertEquals(1, userTags.size());
+        Assert.assertEquals(0, userTags.size());
 
         // set default tag for root
         String setPropStr = "set property for 'root' 'resource_tags.location' = 'default';";