hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Anoop Sam John (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-14970) Backport HBASE-13082 and its sub-jira to branch-1
Date Fri, 18 Dec 2015 07:33:46 GMT

    [ https://issues.apache.org/jira/browse/HBASE-14970?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15063639#comment-15063639
] 

Anoop Sam John commented on HBASE-14970:
----------------------------------------

One point to be checked.
Now we have a Chore Service thread running per region which by default checks for possible
clean up (of compacted away HFiles) in every 2 mins. Now we have a thread pool in HStore level
(with chore threads= 10) and close of each file is handled by one thread.  DO we really need
these many threads in the system?  The max possible #compacted away files per region, can
it be handled by that same chore service thread ?

If we have 100 regions and 2 CFs per region, then we will have 200*10 = 2000 threads??  

> Backport HBASE-13082 and its sub-jira to branch-1
> -------------------------------------------------
>
>                 Key: HBASE-14970
>                 URL: https://issues.apache.org/jira/browse/HBASE-14970
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 1.3.0
>            Reporter: ramkrishna.s.vasudevan
>            Assignee: ramkrishna.s.vasudevan
>         Attachments: HBASE-13082-branch-1.patch, HBASE-14970_branch-1.patch, HBASE-14970_branch-1.patch,
HBASE-14970_branch-1.patch, HBASE-14970_branch-1_1.patch
>
>




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

Mime
View raw message