You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Michael Stack (Jira)" <ji...@apache.org> on 2020/04/20 23:02:00 UTC

[jira] [Resolved] (HBASE-24220) Allow that zk NOTEMPTY multi exception is retryable by running in-series

     [ https://issues.apache.org/jira/browse/HBASE-24220?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Stack resolved HBASE-24220.
-----------------------------------
    Fix Version/s: 2.3.0
                   3.0.0
       Resolution: Fixed

Pushed this patch I've been running locally out to branch-2.3+ given nightlies failed w/ the NOTEMPTY issue in  TestHBaseFsckCleanReplicationBarriers

> Allow that zk NOTEMPTY multi exception is retryable by running in-series
> ------------------------------------------------------------------------
>
>                 Key: HBASE-24220
>                 URL: https://issues.apache.org/jira/browse/HBASE-24220
>             Project: HBase
>          Issue Type: Bug
>          Components: flakies
>            Reporter: Michael Stack
>            Assignee: Michael Stack
>            Priority: Major
>             Fix For: 3.0.0, 2.3.0
>
>         Attachments: 0001-HBASE-24220-Allow-that-zk-NOTEMPTY-multi-exception-i.patch
>
>
> It looks like our zk behavior changed slightly on upgrade. In tests on occasion, I see a fail because the zk delete of dir fails because dir is not empty. Or, this flakey type we just ignored in the past (its rare enough -- I'm just seeing it of late because I'm running tests back-to-back on 4 different hardwares). Let me try allowing NOTEMPTY as a retryable. Lets see how it does.
> Threw in some debug while at it for other issues I've seen in test.



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