[Enhancement] Optimize quorum publish, wait some time util all backen…

…ds finish
StarRocks · May 19, 2022 · 3115820 · 3115820
1 parent 6925939
commit 3115820
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 1 deletion.
diff --git a/fe/fe-core/src/main/java/com/starrocks/common/Config.java b/fe/fe-core/src/main/java/com/starrocks/common/Config.java
@@ -1376,4 +1376,10 @@ public class Config extends ConfigBase {
      */
     @ConfField(mutable = true)
     public static long min_routine_load_lag_for_metrics = 10000;
+
+    /**
+     * after wait quorom_publish_wait_time_ms, will do quorum publish
+     */
+    @ConfField(mutable = true)
+    public static int quorom_publish_wait_time_ms = 500;
 }
diff --git a/fe/fe-core/src/main/java/com/starrocks/transaction/DatabaseTransactionMgr.java b/fe/fe-core/src/main/java/com/starrocks/transaction/DatabaseTransactionMgr.java
@@ -644,7 +644,7 @@ public boolean canTxnFinished(TransactionState txn, Set<Long> errReplicas, Set<L
             return true;
         }
         db.readLock();
-
+        long currentTs = System.currentTimeMillis();
         try {
             // check each table involved in transaction
             for (TableCommitInfo tableCommitInfo : txn.getIdToTableCommitInfos().values()) {
@@ -670,6 +670,7 @@ public boolean canTxnFinished(TransactionState txn, Set<Long> errReplicas, Set<L
 
                     List<MaterializedIndex> allIndices = txn.getPartitionLoadedTblIndexes(tableId, partition);
                     int quorumNum = partitionInfo.getQuorumNum(partitionId);
+                    int replicaNum = partitionInfo.getReplicationNum(partitionId);
                     for (MaterializedIndex index : allIndices) {
                         for (Tablet tablet : index.getTablets()) {
                             int successHealthyReplicaNum = 0;
@@ -705,6 +706,15 @@ public boolean canTxnFinished(TransactionState txn, Set<Long> errReplicas, Set<L
                             if (successHealthyReplicaNum < quorumNum) {
                                 return false;
                             }
+                            // quorum publish will make table unstable
+                            // so that we wait quorom_publish_wait_time_ms util all backend publish finish
+                            // before quorum publish
+                            if (successHealthyReplicaNum != replicaNum
+                                    && !unfinishedBackends.isEmpty()
+                                    && currentTs
+                                            - txn.getCommitTime() < Config.quorom_publish_wait_time_ms) {
+                                return false;
+                            }
                         }
                     }
                 }