[
https://issues.apache.org/jira/browse/HDFS-15108?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17012130#comment-17012130
]
Ayush Saxena commented on HDFS-15108:
-------------------------------------
In {{RouterRpcClient#InvokeMethod}} L425, if failover occurs, the namenode
other than the know ACTIVE serves the request, then that namenode is updated as
ACTIVE by {{namenodeResolver.updateActiveNamenode(nsId, address);}}, but in the
subsequent calls, when {{getNamenodesForNameserviceId}} is called again the
wrong ACTIVE is returned and again this procedure repeats as this method gets
entry from cacheNs as {{cacheNS.get(nsId);}}
> RBF: MembershipNamenodeResolver should invalidate cache incase of active
> namenode update
> ----------------------------------------------------------------------------------------
>
> Key: HDFS-15108
> URL: https://issues.apache.org/jira/browse/HDFS-15108
> Project: Hadoop HDFS
> Issue Type: Bug
> Reporter: Ayush Saxena
> Assignee: Ayush Saxena
> Priority: Major
> Attachments: HDFS-15108-01.patch, HDFS-15108-02.patch
>
>
> If a failover happens, {{namenodeResolver.updateActiveNamenode(nsId,
> address);}} is called, but this doesn't invalidates the cache, so as the next
> time the correct active is fetched.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
