You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ratis.apache.org by "Shashikant Banerjee (Jira)" <ji...@apache.org> on 2019/12/12 12:32:00 UTC
[jira] [Commented] (RATIS-780) Pipeline reports
LeaderNotReadyException after written a bunch of data
[ https://issues.apache.org/jira/browse/RATIS-780?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16994608#comment-16994608 ]
Shashikant Banerjee commented on RATIS-780:
-------------------------------------------
After looking at the heapdump, the issue seems to be arising because for a write chunk request, the corresponding log index never gets updated in the applyTransaction completion map as a result of which the applied index of containerStateMachine never gets updated any further. This seems to be a ozone bug rather than Ratis. Need some more analysis on why the index never gets updated even though the stateMachine is healthy and functioning.
> Pipeline reports LeaderNotReadyException after written a bunch of data
> ----------------------------------------------------------------------
>
> Key: RATIS-780
> URL: https://issues.apache.org/jira/browse/RATIS-780
> Project: Ratis
> Issue Type: Bug
> Reporter: Sammi Chen
> Priority: Blocker
> Attachments: client.log, leader-metrics.log
>
>
> The pipeline failed to serve write request after written a bunch of data.
> There is no WARN or ERROR messags in datanode log file.
> Client log attached.
> Leader node metrids attached.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)