hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-4926) DebugFilter; counts how many KVs skipped before result found, counts storefiles we tripped over
Date Fri, 31 Aug 2012 06:24:07 GMT

    [ https://issues.apache.org/jira/browse/HBASE-4926?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13445714#comment-13445714
] 

stack commented on HBASE-4926:
------------------------------

Transparent would be better than having to add a filter.  Set a flag on Scan and Get?  Hard
part would be adding the info unless we did something like add it as meta on Result returned?
 Should really be in envelope of the response rpc.
                
> DebugFilter; counts how many KVs skipped before result found, counts storefiles we tripped
over
> -----------------------------------------------------------------------------------------------
>
>                 Key: HBASE-4926
>                 URL: https://issues.apache.org/jira/browse/HBASE-4926
>             Project: HBase
>          Issue Type: Task
>            Reporter: stack
>
> I'm not sure this possible but chatting w/ J-D after he finished helping a crew who had
a slow scan, we need ways to help folks figure out why stuff is slow on the server side. 
In this case, the user had a region that had gigs of deletes in it that had to be skipped
before the scanner got to a value to return so scans looked slow though working 'correctly'.
 If the user could enable a DebugFilter than did counts of what Scans were skipping and size
of data skipped among other things, it might help w/ the situation above -- and perhaps with
others (What else could we count in a filter?)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message