(iotdb) branch master updated: [IOTDB-6309] Lager interval for auto ratis leader balance (#12116)

[tanxinyu]

This is an automated email from the ASF dual-hosted git repository.

tanxinyu pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/iotdb.git


The following commit(s) were added to refs/heads/master by this push:
     new 46e47672c40 [IOTDB-6309] Lager interval for auto ratis leader balance 
(#12116)
46e47672c40 is described below

commit 46e47672c403bc8d363588a8856586e56b1dd2ad
Author: Yongzao <yong...@apache.org>
AuthorDate: Fri Mar 8 16:32:11 2024 +0800

    [IOTDB-6309] Lager interval for auto ratis leader balance (#12116)
---
 .../manager/load/balancer/RouteBalancer.java       | 24 ++++++++++++++++++++--
 1 file changed, 22 insertions(+), 2 deletions(-)

diff --git 
a/iotdb-core/confignode/src/main/java/org/apache/iotdb/confignode/manager/load/balancer/RouteBalancer.java
 
b/iotdb-core/confignode/src/main/java/org/apache/iotdb/confignode/manager/load/balancer/RouteBalancer.java
index 463a5df0dc7..1f3fb4efb55 100644
--- 
a/iotdb-core/confignode/src/main/java/org/apache/iotdb/confignode/manager/load/balancer/RouteBalancer.java
+++ 
b/iotdb-core/confignode/src/main/java/org/apache/iotdb/confignode/manager/load/balancer/RouteBalancer.java
@@ -86,6 +86,9 @@ public class RouteBalancer {
           // The simple consensus protocol will always automatically designate 
itself as the leader
           || 
ConsensusFactory.SIMPLE_CONSENSUS.equals(SCHEMA_REGION_CONSENSUS_PROTOCOL_CLASS);
 
+  // The interval of retrying to balance ratis leader after the last failed 
time
+  private static final long BALANCE_RATIS_LEADER_FAILED_INTERVAL = 60 * 1000L;
+
   private final IManager configManager;
 
   /** RegionRouteMap */
@@ -94,6 +97,9 @@ public class RouteBalancer {
   // For generating optimal RegionPriorityMap
   private final IPriorityBalancer priorityRouter;
 
+  private long lastFailedTimeForBalanceRatisSchemaLeader = 0;
+  private long lastFailedTimeForBalanceRatisDataLeader = 0;
+
   public RouteBalancer(IManager configManager) {
     this.configManager = configManager;
 
@@ -127,10 +133,16 @@ public class RouteBalancer {
   public synchronized Map<TConsensusGroupId, Pair<Integer, Integer>> 
balanceRegionLeader() {
     Map<TConsensusGroupId, Pair<Integer, Integer>> differentRegionLeaderMap =
         new ConcurrentHashMap<>();
-    if (IS_ENABLE_AUTO_LEADER_BALANCE_FOR_SCHEMA_REGION) {
+    if (IS_ENABLE_AUTO_LEADER_BALANCE_FOR_SCHEMA_REGION
+        && 
(!ConsensusFactory.RATIS_CONSENSUS.equals(SCHEMA_REGION_CONSENSUS_PROTOCOL_CLASS)
+            || System.currentTimeMillis() - 
lastFailedTimeForBalanceRatisSchemaLeader
+                > BALANCE_RATIS_LEADER_FAILED_INTERVAL)) {
       
differentRegionLeaderMap.putAll(balanceRegionLeader(TConsensusGroupType.SchemaRegion));
     }
-    if (IS_ENABLE_AUTO_LEADER_BALANCE_FOR_DATA_REGION) {
+    if (IS_ENABLE_AUTO_LEADER_BALANCE_FOR_DATA_REGION
+        && 
(!ConsensusFactory.RATIS_CONSENSUS.equals(DATA_REGION_CONSENSUS_PROTOCOL_CLASS)
+            || System.currentTimeMillis() - 
lastFailedTimeForBalanceRatisDataLeader
+                > BALANCE_RATIS_LEADER_FAILED_INTERVAL)) {
       
differentRegionLeaderMap.putAll(balanceRegionLeader(TConsensusGroupType.DataRegion));
     }
 
@@ -199,6 +211,14 @@ public class RouteBalancer {
                   
clientHandler.getRequest(i).getNewLeaderNode().getDataNodeId());
         } else {
           
differentRegionLeaderMap.remove(clientHandler.getRequest(i).getRegionId());
+          if (TConsensusGroupType.SchemaRegion.equals(regionGroupType)
+              && 
ConsensusFactory.RATIS_CONSENSUS.equals(SCHEMA_REGION_CONSENSUS_PROTOCOL_CLASS))
 {
+            lastFailedTimeForBalanceRatisSchemaLeader = 
System.currentTimeMillis();
+          }
+          if (TConsensusGroupType.DataRegion.equals(regionGroupType)
+              && 
ConsensusFactory.RATIS_CONSENSUS.equals(DATA_REGION_CONSENSUS_PROTOCOL_CLASS)) {
+            lastFailedTimeForBalanceRatisDataLeader = 
System.currentTimeMillis();
+          }
           LOGGER.error(
               "[LeaderBalancer] Failed to change the leader of Region: {} to 
DataNode: {}",
               clientHandler.getRequest(i).getRegionId(),