[jira] [Commented] (HDFS-15023) [SBN read] ZKFC should check the state before joining the election

["Fei Hui (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HDFS-15023?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16987578#comment-16987578 ] 

Fei Hui commented on HDFS-15023:
--------------------------------

[~ayushtkn] Thanks for your comments. I had a mistake. I shouldn't change the fix mentioned on HDFS-14961
Upload v002 patch with UT 

> [SBN read] ZKFC should check the state before joining the election
> ------------------------------------------------------------------
>
>                 Key: HDFS-15023
>                 URL: https://issues.apache.org/jira/browse/HDFS-15023
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>            Reporter: Fei Hui
>            Assignee: Fei Hui
>            Priority: Major
>         Attachments: HDFS-15023.001.patch
>
>
> As discussed HDFS-14961, ZKFC should not join election when its state is observer.
> Right now when namemode was an observer,  it joined election and it would be become a standby.
> MonitorDaemon thread callchain is that
> doHealthChecks -> enterState(State.SERVICE_HEALTHY) -> recheckElectability() -> elector.joinElection(targetToData(localTarget)) -> joinElectionInternal -> createLockNodeAsync
> callBack for zookeeper
> processResult -> becomeStandby



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org