[jira] [Resolved] (HBASE-25643) The delayed FlushRegionEntry should be removed when we need a non-delayed one

["Michael Stack (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-25643?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Stack resolved HBASE-25643.
-----------------------------------
    Fix Version/s: 2.5.0
                   3.0.0-alpha-1
     Hadoop Flags: Reviewed
       Resolution: Fixed

Merged to branch-2+. Thanks for the nice fix [~filtertip] It would not go back to branch-2.4 so if you'd like it to go to there, please make a backport subtask and attach a new PR please.
Thanks for review [~anoop.hbase]

> The delayed FlushRegionEntry should be removed when we need a non-delayed one
> -----------------------------------------------------------------------------
>
>                 Key: HBASE-25643
>                 URL: https://issues.apache.org/jira/browse/HBASE-25643
>             Project: HBase
>          Issue Type: Improvement
>          Components: regionserver
>            Reporter: Zheng Wang
>            Assignee: Zheng Wang
>            Priority: Major
>             Fix For: 3.0.0-alpha-1, 2.5.0
>
>
> The regionserver periodically check all the regions, if one not flushed for long time, then it will create a delayed FlushRegionEntry, the delay range is 0~300s.
> During the delay time, if many data are written to the region suddenly, we can not do the flush immediately due to the existing one in regionsInQueue, then the RegionTooBusyException will occur.
> It is better to improve the logic here, that the delayed entry should be replaced by the non-delayed one.



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