HBASE-29804: NullPointerException in WorkerAssigner.serverAdded during Master Shutdown and Restart

hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java

@@ -1821,6 +1821,14 @@ protected void stopServiceThreads() {
     }
 
     LOG.debug("Stopping service threads");
+
+    // Stop SplitWALManager before procedure executor to unregister its WorkerAssigner from
+    // ServerManager and prevent NPE when serverAdded() is called after procedureExecutor is null.
+    // See HBASE-29804.
+    if (this.splitWALManager != null) {
+      this.splitWALManager.stop();
+    }
+
     // stop procedure executor prior to other services such as server manager and assignment
     // manager, as these services are important for some running procedures. See HBASE-24117 for
     // example.

hbase-server/src/main/java/org/apache/hadoop/hbase/master/SplitWALManager.java

@@ -176,4 +176,13 @@ public void releaseSplitWALWorker(ServerName worker) {
   public void addUsedSplitWALWorker(ServerName worker) {
     splitWorkerAssigner.addUsedWorker(worker);
   }
+
+  /**
+   * Stop the SplitWALManager and unregister the WorkerAssigner from ServerManager. This should be
+   * called during shutdown to prevent NPE when serverAdded() is triggered after procedureExecutor
+   * is set to null.
+   */
+  public void stop() {
+    splitWorkerAssigner.stop();
+  }
 }

hbase-server/src/main/java/org/apache/hadoop/hbase/master/WorkerAssigner.java

@@ -23,8 +23,10 @@
 import java.util.Map;
 import java.util.Optional;
 import org.apache.hadoop.hbase.ServerName;
+import org.apache.hadoop.hbase.master.procedure.MasterProcedureEnv;
 import org.apache.hadoop.hbase.procedure2.Procedure;
 import org.apache.hadoop.hbase.procedure2.ProcedureEvent;
+import org.apache.hadoop.hbase.procedure2.ProcedureExecutor;
 import org.apache.hadoop.hbase.procedure2.ProcedureSuspendedException;
 import org.apache.yetus.audience.InterfaceAudience;
 
@@ -80,7 +82,13 @@ public synchronized void release(ServerName serverName) {
   @Override
   public synchronized void serverAdded(ServerName worker) {
     if (!event.isReady()) {
-      event.wake(master.getMasterProcedureExecutor().getEnvironment().getProcedureScheduler());
+      ProcedureExecutor<MasterProcedureEnv> executor = master.getMasterProcedureExecutor();
+      if (executor != null) {
+        MasterProcedureEnv env = executor.getEnvironment();
+        if (env != null) {
+          event.wake(env.getProcedureScheduler());
+        }
+      }
     }
   }
 
@@ -93,4 +101,15 @@ public synchronized void addUsedWorker(ServerName worker) {
   public Integer getAvailableWorker(ServerName serverName) {
     return currentWorkers.get(serverName);
   }
+
+  /**
+   * Stop the WorkerAssigner and unregister it from ServerManager. This should be called during
+   * shutdown to prevent NPE when serverAdded() is triggered after procedureExecutor is set to null.
+   */
+  public void stop() {
+    ServerManager sm = this.master.getServerManager();
+    if (sm != null) {
+      sm.unregisterListener(this);
+    }
+  }
 }

hbase-server/src/main/java/org/apache/hadoop/hbase/master/snapshot/SnapshotManager.java

@@ -1255,6 +1255,11 @@ public void stop(String why) {
     } catch (IOException e) {
       LOG.error("stop ProcedureCoordinator error", e);
     }
+    // Unregister the WorkerAssigner from ServerManager to prevent NPE during shutdown
+    // when serverAdded() is called after procedureExecutor is set to null
+    if (verifyWorkerAssigner != null) {
+      verifyWorkerAssigner.stop();
+    }
   }
 
   @Override