HBASE-27311 The implementation of syncReplicationPeerLock is incomple…

…te (#4715) Signed-off-by: Xin Sun <[email protected]>
apache · Aug 23, 2022 · 950ad8d · 950ad8d
1 parent 00a719e
commit 950ad8d
Show file tree

Hide file tree

Showing 3 changed files with 25 additions and 5 deletions.
diff --git a/hbase-server/src/main/java/org/apache/hadoop/hbase/master/replication/AddPeerProcedure.java b/hbase-server/src/main/java/org/apache/hadoop/hbase/master/replication/AddPeerProcedure.java
@@ -23,6 +23,7 @@
 import org.apache.hadoop.hbase.master.procedure.MasterProcedureEnv;
 import org.apache.hadoop.hbase.master.procedure.ProcedurePrepareLatch;
 import org.apache.hadoop.hbase.procedure2.ProcedureStateSerializer;
+import org.apache.hadoop.hbase.procedure2.ProcedureSuspendedException;
 import org.apache.hadoop.hbase.replication.ReplicationException;
 import org.apache.hadoop.hbase.replication.ReplicationPeerConfig;
 import org.apache.yetus.audience.InterfaceAudience;
@@ -91,13 +92,18 @@ protected void releaseLatch(MasterProcedureEnv env) {
 
   @Override
   protected void prePeerModification(MasterProcedureEnv env)
-    throws IOException, ReplicationException, InterruptedException {
+    throws IOException, ReplicationException, ProcedureSuspendedException {
     MasterCoprocessorHost cpHost = env.getMasterCoprocessorHost();
     if (cpHost != null) {
       cpHost.preAddReplicationPeer(peerId, peerConfig);
     }
     if (peerConfig.isSyncReplication()) {
-      env.getReplicationPeerManager().acquireSyncReplicationPeerLock();
+      if (!env.getReplicationPeerManager().tryAcquireSyncReplicationPeerLock()) {
+        throw suspend(env.getMasterConfiguration(),
+          backoff -> LOG.warn(
+            "Can not acquire sync replication peer lock for peer {}, sleep {} secs", peerId,
+            backoff / 1000));
+      }
     }
     env.getReplicationPeerManager().preAddPeer(peerId, peerConfig);
   }
@@ -119,6 +125,20 @@ protected void postPeerModification(MasterProcedureEnv env)
     }
   }
 
+  @Override
+  protected void afterReplay(MasterProcedureEnv env) {
+    if (getCurrentState() == getInitialState()) {
+      // will try to acquire the lock when executing the procedure, no need to acquire it here
+      return;
+    }
+    if (peerConfig.isSyncReplication()) {
+      if (!env.getReplicationPeerManager().tryAcquireSyncReplicationPeerLock()) {
+        throw new IllegalStateException(
+          "Can not acquire sync replication peer lock for peer " + peerId);
+      }
+    }
+  }
+
   @Override
   protected void serializeStateData(ProcedureStateSerializer serializer) throws IOException {
     super.serializeStateData(serializer);

diff --git a/...-server/src/main/java/org/apache/hadoop/hbase/master/replication/ModifyPeerProcedure.java b/...-server/src/main/java/org/apache/hadoop/hbase/master/replication/ModifyPeerProcedure.java
@@ -60,7 +60,7 @@ protected ModifyPeerProcedure(String peerId) {
    * all checks passes then the procedure can not be rolled back any more.
    */
   protected abstract void prePeerModification(MasterProcedureEnv env)
-    throws IOException, ReplicationException, InterruptedException;
+    throws IOException, ReplicationException, ProcedureSuspendedException;
 
   protected abstract void updatePeerStorage(MasterProcedureEnv env) throws ReplicationException;
 

diff --git a/...rver/src/main/java/org/apache/hadoop/hbase/master/replication/ReplicationPeerManager.java b/...rver/src/main/java/org/apache/hadoop/hbase/master/replication/ReplicationPeerManager.java
@@ -587,8 +587,8 @@ private boolean isStringEquals(String s1, String s2) {
     return s1.equals(s2);
   }
 
-  public void acquireSyncReplicationPeerLock() throws InterruptedException {
-    syncReplicationPeerLock.acquire();
+  public boolean tryAcquireSyncReplicationPeerLock() {
+    return syncReplicationPeerLock.tryAcquire();
   }
 
   public void releaseSyncReplicationPeerLock() {