Jane He created CASSANDRA-21641:
-----------------------------------
Summary: Duplicate NativeTransportService.stop() calls in
CassandraDaemon.destroyClientTransports()
Key: CASSANDRA-21641
URL: https://issues.apache.org/jira/browse/CASSANDRA-21641
Project: Apache Cassandra
Issue Type: Bug
Reporter: Jane He
Assignee: Jane He
CassandraDaemon.[destroyClientTransports()|https://github.com/apache/cassandra/blob/b20985d744984304434d2713faa59b4a6f766174/src/java/org/apache/cassandra/service/CassandraDaemon.java#L813-L818]
first calls {{{}stopNativeTransport(){}}}, which is essentially
{{{}nativeTransportService.stop(false){}}}, then calls
{{{}nativeTransportService.destroy(){}}}, which internally calls
{{nativeTransportService.stop(false)}} , too. So it calls
{{nativeTransportService.stop(false)}} twice in a row. I believe it's
introduced from [this merge
commit|https://github.com/apache/cassandra/commit/42989cee287ddacf30db331829bf927695816966#diff-957f2fa6365cb92f19b74347fee7a9f310a07e32c3112f35196dc17462ec7269].
Although stopping the transport service seems to be idempotent, so stopping it
twice didn't surface as a bug, still I think we should remove the redundant
{{stopNativeTransport()}} call, to avoid confusion and the risks of doubling
the grace period.
As this will be a one-line change, I plan to fix this as a part of the
CASSANDRA-21191 [PR|https://github.com/apache/cassandra/pull/4953].
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
