-
Type:
Bug
-
Resolution: Fixed
-
Priority:
Major - P3
-
Affects Version/s: None
-
Component/s: None
-
Labels:None
-
Fully Compatible
-
ALL
-
140
The source of the issue appears to be improper shutdown of EventsPublisher/SingleServerDiscoveryMonitor. Following a similar logic to StreamableReplicaSetMonitor should enable us to gracefully shutdown these components.
[js_test:shard_split_basic_test] d20270| 2022-03-18T15:16:15.247+00:00 I - 4333222 [ShardSplitDonorService-3] "RSM received error response","attr":{"host":"ip-10-122-17-171.ec2.internal:20275","error":"ShutdownInProgress: Shutdown in progress","replicaSet":"","response":{}} [js_test:shard_split_basic_test] d20270| 2022-03-18T15:16:15.247+00:00 F - 5106800 [ShardSplitDonorService-3] "Theoretical deadlock found on use of latch","attr":{"reason":"Latch acquired after other latch of lower level","latch":{"name":"TopologyEventsPublisher::_eventQueueMutex","latchId":11855,"level":6,"file":"src/mongo/client/sdam/topology_listener.h","line":99},"latchesHeld":[{"name":"SingleServerDiscoveryMonitor::mutex","latchId":11858,"level":4,"file":"src/mongo/client/server_discovery_monitor.cpp","line":85}]} [js_test:shard_split_basic_test] d20270| 2022-03-18T15:16:15.247+00:00 F ASSERT 23089 [ShardSplitDonorService-3] "Fatal assertion","attr":{"msgid":5106800,"file":"src/mongo/util/latch_analyzer.cpp","line":229} [js_test:shard_split_basic_test] d20270| 2022-03-18T15:16:15.247+00:00 F ASSERT 23090 [ShardSplitDonorService-3] "\n\n***aborting after fassert() failure\n\n" [js_test:shard_split_basic_test] d20270| 2022-03-18T15:16:15.247+00:00 F CONTROL 4757800 [ShardSplitDonorService-3] "Writing fatal message","attr":{"message":"Got signal: 6 (Aborted).\n"}