Treat TransportService stopped error as node is closing (#39800)

If TransportService is stopped before a shard-failure request is sent but after the request is registered, TransportService will notify ReplicationOperation a TransportException with an error message: "transport stop, action: internal:cluster/shard/failure". Relates #39584
elastic · Mar 8, 2019 · 027f210 · 027f210
1 parent d7aeb02
commit 027f210
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 2 deletions.
diff --git a/server/src/main/java/org/elasticsearch/action/support/replication/ReplicationOperation.java b/server/src/main/java/org/elasticsearch/action/support/replication/ReplicationOperation.java
@@ -206,7 +206,9 @@ public String toString() {
     private void onNoLongerPrimary(Exception failure) {
         final Throwable cause = ExceptionsHelper.unwrapCause(failure);
         final boolean nodeIsClosing = cause instanceof NodeClosedException
-            || (cause instanceof TransportException && "TransportService is closed stopped can't send request".equals(cause.getMessage()));
+            || (cause instanceof TransportException &&
+                ("TransportService is closed stopped can't send request".equals(cause.getMessage())
+                || "transport stopped, action: internal:cluster/shard/failure".equals(cause.getMessage())));
         final String message;
         if (nodeIsClosing) {
             message = String.format(Locale.ROOT,

diff --git a/...src/test/java/org/elasticsearch/action/support/replication/ReplicationOperationTests.java b/...src/test/java/org/elasticsearch/action/support/replication/ReplicationOperationTests.java
@@ -205,8 +205,11 @@ public void testNoLongerPrimary() throws Exception {
             shardActionFailure = new NodeClosedException(new DiscoveryNode("foo", buildNewFakeTransportAddress(), Version.CURRENT));
         } else if (randomBoolean()) {
             shardActionFailure = new SendRequestTransportException(
-                new DiscoveryNode("foo", buildNewFakeTransportAddress(), Version.CURRENT), "internal:cluster/shard/failure",
+                new DiscoveryNode("foo", buildNewFakeTransportAddress(), Version.CURRENT), ShardStateAction.SHARD_FAILED_ACTION_NAME,
                 new TransportException("TransportService is closed stopped can't send request"));
+        } else if (randomBoolean()) {
+            shardActionFailure = new TransportException(
+                "transport stopped, action: " + ShardStateAction.SHARD_FAILED_ACTION_NAME);
         } else {
             shardActionFailure = new ShardStateAction.NoLongerPrimaryShardException(failedReplica.shardId(), "the king is dead");
         }