cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-3444) Secondary Index doesn't clean up indexed CFS on remove and Streaming test failure.
Date Sun, 06 Nov 2011 02:02:51 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-3444?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13144886#comment-13144886
] 

Jonathan Ellis commented on CASSANDRA-3444:
-------------------------------------------

I think this is showing that SIM needs some additional refactoring -- currently removeIndex
is poorly named, because it leaves the index metadata around (which this patch tries to address)
but invalidates the CFS representing it.  This is addressed by 3437 v2 as well.  (May have
been addressed by v1, but I'm sure it is by v2. :)

So I would prefer to close this in favor of 3437.
                
> Secondary Index doesn't clean up indexed CFS on remove and Streaming test failure.
> ----------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-3444
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3444
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.0.0
>            Reporter: Sylvain Lebresne
>            Assignee: Sylvain Lebresne
>            Priority: Trivial
>             Fix For: 1.0.3
>
>         Attachments: 0001-clean-indexes-post-remove-and-streaming-tests.patch
>
>
> The initial reason for that issue is because StreamingTransferTest is broken in trunk.
It has been broken by CASSANDRA-3116 because the latter is too efficient. More precisely StreamingTransferTest
create a CF with an index, then it calls unreferenceSSTables() on that CF to remove all sstable,
and then try a transfer (that recreate the file and index basically). But when unreferenceSSTables()
is called, it does fully remove the indexes in that the CFS object for the indexes stays.
Post CASSANDRA-3116, this is problem because that CFS has been invalidated and thus nothing
can be added back to it.
> Long story short, I believe that the fact that SecondaryIndexManager.removeAllIndex doesn't
really unreference the CFS objects is not expected in the first place. The patch fixes that
and update the StreamingTransferTest accordingly (fixing it as far as trunk is concerned).

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message