hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jim Kellerman (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-955) Stop HLog thread before starting a new one
Date Mon, 27 Oct 2008 21:15:44 GMT

    [ https://issues.apache.org/jira/browse/HBASE-955?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12643062#action_12643062
] 

Jim Kellerman commented on HBASE-955:
-------------------------------------

Will be not applicable after upcoming patch to HBASE-728, because HLog will no longer be either
a Thread or a Chore. Instead a HLog object will manage the HLog and nothing else.

LogRoller, (a Thread) and LogFlusher (a Chore) will be responsible for getting the HLog to
disk.

> Stop HLog thread before starting a new one
> ------------------------------------------
>
>                 Key: HBASE-955
>                 URL: https://issues.apache.org/jira/browse/HBASE-955
>             Project: Hadoop HBase
>          Issue Type: Sub-task
>          Components: regionserver
>    Affects Versions: 0.19.0
>            Reporter: Jim Kellerman
>            Assignee: Jim Kellerman
>            Priority: Blocker
>             Fix For: 0.19.0
>
>
> Currently, when we encounter a MSG_CALL_SERVER_STARTUP situation in the master the region
server discards the old HLog for a new one. However, it does not stop the thread associated
with the old HLog. Thus we have an orphaned thread doing nothing, except consuming CPU. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message