hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Elliott Clark (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-14712) MasterProcWALs never clean up
Date Wed, 28 Oct 2015 21:42:27 GMT

    [ https://issues.apache.org/jira/browse/HBASE-14712?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14979298#comment-14979298
] 

Elliott Clark commented on HBASE-14712:
---------------------------------------

[~mbertozzi] You around to look at this ?

CI cluster running 1.2.0-SNAPSHOT has 120k of master logs dating back about a week since the
last time it was cleaned up.

Each time a master tries to become active it has that many different logs to recover lease
on, and read. This ends up ddosing the namenode. It runs out of tcp buffer space and everything
falls over.

> MasterProcWALs never clean up
> -----------------------------
>
>                 Key: HBASE-14712
>                 URL: https://issues.apache.org/jira/browse/HBASE-14712
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Elliott Clark
>            Priority: Blocker
>
> MasterProcWALs directory grows pretty much un-bounded. Because of that when master failover
happens the NN is flooded with connections and everything grinds to a halt.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message