[
https://issues.apache.org/jira/browse/HADOOP-18624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17693039#comment-17693039
]
ASF GitHub Bot commented on HADOOP-18624:
-----------------------------------------
ZanderXu commented on code in PR #5367:
URL: https://github.com/apache/hadoop/pull/5367#discussion_r1116588469
##########
hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/ipc/Client.java:
##########
@@ -1485,6 +1487,10 @@ Writable call(RPC.RpcKind rpcKind, Writable rpcRequest,
releaseAsyncCall();
}
throw e;
+ } finally {
+ if (!success) {
+ connection.calls.remove(call.id);
Review Comment:
Thanks @xinglin for your comment.
We don't need to remove the call from `rpcRequestQueue`, because the
`rpcRequestThread` will poll it and send to the connection.
> Leaked calls may cause ObserverNameNode OOM.
> --------------------------------------------
>
> Key: HADOOP-18624
> URL: https://issues.apache.org/jira/browse/HADOOP-18624
> Project: Hadoop Common
> Issue Type: Bug
> Reporter: ZanderXu
> Assignee: ZanderXu
> Priority: Major
> Labels: pull-request-available
>
> Leaked calls may cause ObserverNameNode OOM.
>
> During Observer Namenode tailing edits from JournalNode, it will cancel slow
> request with an interruptException if there are a majority of successful
> responses.
> There is a bug in Client.java, it will not clean the interrupted call from
> the calls. The leaked calls may cause ObserverNameNode OOM.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
