Improve exception handling on TransportMasterNodeAction (#29314)

We have seen exceptions bubble up to the uncaught exception handler. Checking the blocks can
lead for example to IndexNotFoundException when the indices are resolved. In order to make
TransportMasterNodeAction more resilient against such expected exceptions, this code change
wraps the execution of doStart() into a try catch and informs the listener in case of failures.

server/src/main/java/org/elasticsearch/action/support/master/TransportMasterNodeAction.java

@@ -145,6 +145,7 @@ public abstract class TransportMasterNodeAction<Request extends MasterNodeReques
         }
 
         protected void doStart(ClusterState clusterState) {
+            try {
                 final Predicate<ClusterState> masterChangePredicate = MasterNodeChangePredicate.build(clusterState);
                 final DiscoveryNodes nodes = clusterState.nodes();
                 if (nodes.isLocalNodeElectedMaster() || localExecute(request)) {
@@ -156,8 +157,14 @@ public abstract class TransportMasterNodeAction<Request extends MasterNodeReques
                         } else {
                             logger.trace("can't execute due to a cluster block, retrying", blockException);
                             retry(blockException, newState -> {
+                                try {
                                     ClusterBlockException newException = checkBlock(request, newState);
                                     return (newException == null || !newException.retryable());
+                                } catch (Exception e) {
+                                    // accept state as block will be rechecked by doStart() and listener.onFailure() then called
+                                    logger.trace("exception occurred during cluster block checking, accepting state", e);
+                                    return true;
+                                }
                             });
                         }
                     } else {
@@ -209,6 +216,9 @@ public abstract class TransportMasterNodeAction<Request extends MasterNodeReques
                         });
                     }
                 }
+            } catch (Exception e) {
+                listener.onFailure(e);
+            }
         }
 
         private void retry(final Throwable failure, final Predicate<ClusterState> statePredicate) {

server/src/test/java/org/elasticsearch/action/support/master/TransportMasterNodeActionTests.java

@@ -242,6 +242,39 @@ public class TransportMasterNodeActionTests extends ESTestCase {
         }
     }
 
+    public void testCheckBlockThrowsException() throws InterruptedException {
+        boolean throwExceptionOnRetry = randomBoolean();
+        Request request = new Request().masterNodeTimeout(TimeValue.timeValueSeconds(60));
+        PlainActionFuture<Response> listener = new PlainActionFuture<>();
+
+        ClusterBlock block = new ClusterBlock(1, "", true, true,
+            false, randomFrom(RestStatus.values()), ClusterBlockLevel.ALL);
+        ClusterState stateWithBlock = ClusterState.builder(ClusterStateCreationUtils.state(localNode, localNode, allNodes))
+            .blocks(ClusterBlocks.builder().addGlobalBlock(block)).build();
+        setState(clusterService, stateWithBlock);
+
+        new Action(Settings.EMPTY, "testAction", transportService, clusterService, threadPool) {
+            @Override
+            protected ClusterBlockException checkBlock(Request request, ClusterState state) {
+                Set<ClusterBlock> blocks = state.blocks().global();
+                if (throwExceptionOnRetry == false || blocks.isEmpty()) {
+                    throw new RuntimeException("checkBlock has thrown exception");
+                }
+                return new ClusterBlockException(blocks);
+
+            }
+        }.execute(request, listener);
+
+        if (throwExceptionOnRetry == false) {
+            assertListenerThrows("checkBlock has thrown exception", listener, RuntimeException.class);
+        } else {
+            assertFalse(listener.isDone());
+            setState(clusterService, ClusterState.builder(ClusterStateCreationUtils.state(localNode, localNode, allNodes))
+                .blocks(ClusterBlocks.EMPTY_CLUSTER_BLOCK).build());
+            assertListenerThrows("checkBlock has thrown exception", listener, RuntimeException.class);
+        }
+    }
+
     public void testForceLocalOperation() throws ExecutionException, InterruptedException {
         Request request = new Request();
         PlainActionFuture<Response> listener = new PlainActionFuture<>();