Fix Flaky Test testPendingTasksWithClusterNotRecoveredBlock

Signed-off-by: kkewwei <[email protected]> Signed-off-by: kkewwei <[email protected]> test
kkewwei · Feb 20, 2025 · ea2e674 · ea2e674
1 parent c34cedc
commit ea2e674
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 3 deletions.
diff --git a/...ernalClusterTest/java/org/opensearch/action/admin/cluster/tasks/PendingTasksBlocksIT.java b/...ernalClusterTest/java/org/opensearch/action/admin/cluster/tasks/PendingTasksBlocksIT.java
@@ -32,6 +32,7 @@
 
 package org.opensearch.action.admin.cluster.tasks;
 
+import org.opensearch.action.admin.cluster.node.info.NodesInfoRequest;
 import org.opensearch.common.settings.Settings;
 import org.opensearch.gateway.GatewayService;
 import org.opensearch.test.InternalTestCluster;
@@ -91,11 +92,12 @@ public void testPendingTasksWithClusterNotRecoveredBlock() throws Exception {
         }
 
         // restart the cluster but prevent it from performing state recovery
-        final int nodeCount = client().admin().cluster().prepareNodesInfo("data:true", "cluster_manager:true").get().getNodes().size();
+        NodesInfoRequest nodesInfoRequest = new NodesInfoRequest();
+        final int dataNodesCount = client().admin().cluster().prepareNodesInfo("data:true").get().getNodes().size();
         internalCluster().fullRestart(new InternalTestCluster.RestartCallback() {
             @Override
             public Settings onNodeStopped(String nodeName) {
-                return Settings.builder().put(GatewayService.RECOVER_AFTER_DATA_NODES_SETTING.getKey(), nodeCount + 1).build();
+                return Settings.builder().put(GatewayService.RECOVER_AFTER_DATA_NODES_SETTING.getKey(), dataNodesCount + 1).build();
             }
 
             @Override

diff --git a/server/src/main/java/org/opensearch/action/support/TransportActions.java b/server/src/main/java/org/opensearch/action/support/TransportActions.java
@@ -67,7 +67,11 @@ public static boolean isReadOverrideException(Exception e) {
     }
 
     public static boolean isRetryableSearchException(final Exception e) {
-        return (OpenSearchException.status(e).getStatus() / 100 != 4) && (e.getCause() instanceof TaskCancelledException == false);
+
+        return (OpenSearchException.status(e).getStatus() / 100 != 4) && (e.getCause() instanceof TaskCancelledException == false)
+            // There exists a scenario where a primary shard (0 replicas) relocates and is in POST_RECOVERY on the
+            // target node but already deleted on the source node. Search request should still work.
+            || (e.getCause() instanceof IndexNotFoundException);
     }
 
 }