hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ramkrishna.s.vasudevan (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-10447) Memstore flusher scans storefiles also when the scanner heap gets reset
Date Tue, 04 Feb 2014 06:14:10 GMT

    [ https://issues.apache.org/jira/browse/HBASE-10447?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13890431#comment-13890431

ramkrishna.s.vasudevan commented on HBASE-10447:

Thanks Lars.
I was about to commit and once verified in 0.96 code.  There we don't add the observers even
while doing compaction.
So should we make it like that in 0.98 and trunk also.  In the sense only for user reads the
observers are added. The current change would any way add observers for compaction also as
it was in 0.98 and trunk.  Would get this clarified once before I could commit it. Suggestions?

> Memstore flusher scans storefiles also when the scanner heap gets reset
> -----------------------------------------------------------------------
>                 Key: HBASE-10447
>                 URL: https://issues.apache.org/jira/browse/HBASE-10447
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 0.98.0, 0.99.0,
>            Reporter: ramkrishna.s.vasudevan
>            Assignee: ramkrishna.s.vasudevan
>            Priority: Blocker
>             Fix For: 0.98.0, 0.96.2, 0.99.0
>         Attachments: HBASE-10447_0.98.patch, HBASE-10447_trunk.patch
> See the mail thread
> http://osdir.com/ml/general/2014-01/msg61294.html
> In case of flush we create a memstore flusher which in turn creates a  StoreScanner backed
by a Single ton MemstoreScanner.  
> But this scanner also registers for any updates in the reader in the HStore.  Is this
> If this happens then any update on the reader may nullify the current heap and the entire
Scanner Stack is reset, but this time with the other scanners for all the files that satisfies
the last top key.  So the flush that happens on the memstore holds the storefile scanners
also in the heap that was recreated but originally the intention was to create a scanner on
the memstore alone.

This message was sent by Atlassian JIRA

View raw message