You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hbase.apache.org by "Wellington Chevreuil (JIRA)" <ji...@apache.org> on 2019/02/01 01:42:00 UTC

[jira] [Commented] (HBASE-18738) Ride over HDFS 'safe mode' post-launch

    [ https://issues.apache.org/jira/browse/HBASE-18738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16757867#comment-16757867 ] 

Wellington Chevreuil commented on HBASE-18738:
----------------------------------------------

So basically accepts reads, but backoff write requests. Internal subsystems should be paused:
* Region Balancer
* AM
* LogRoll
* LogCleaner
* Flusher
* Scheduled Compactions

> Ride over HDFS 'safe mode' post-launch
> --------------------------------------
>
>                 Key: HBASE-18738
>                 URL: https://issues.apache.org/jira/browse/HBASE-18738
>             Project: HBase
>          Issue Type: Improvement
>          Components: Operability
>            Reporter: stack
>            Assignee: Wellington Chevreuil
>            Priority: Major
>             Fix For: 2.2.0
>
>
> A coworker has an integration test that includes setting HDFS into 'safe mode' post-launch of the cluster. He finds that HBase just falls over. We should do better. We should be able to recognize the switch and going into stasis recovering once 'safe mode' has expired.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)