cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Shane (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-4711) Duplicate column names with DynamicCompositeType
Date Fri, 26 Oct 2012 14:37:12 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-4711?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13484957#comment-13484957
] 

Shane commented on CASSANDRA-4711:
----------------------------------

Thanks for the quick response on this.  Do you have a time frame for when 1.1.7 is expected
to be released?  I'm seeing this issue in a production environment and would like to get the
fix in place as quickly as possible.  If the release won't be happening in the very near future,
can you patch this into 1.1.6 as well.  While I can download the source and apply the patch
locally, it'd be my preference not to.
                
> Duplicate column names with DynamicCompositeType
> ------------------------------------------------
>
>                 Key: CASSANDRA-4711
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-4711
>             Project: Cassandra
>          Issue Type: Bug
>    Affects Versions: 1.1.5
>         Environment: CentOS 5.8 x86_64
> Cassandra 1.1.5
> Hector 1.1-2
>            Reporter: Shane
>            Assignee: Sylvain Lebresne
>             Fix For: 1.1.7
>
>         Attachments: 4711.txt, DuplicateColumnName.java
>
>
> I have a column family whose comparator is DynamicCompositeType and validation is CounterColumnType.
 During automated testing, there have been occasions where a counter column is created twice,
throwing off the query results for the column.
> Doing a 'get' via the cli, I see the following output for the row:
> => (counter=s@language:b@00000001:s@pt_BR, value=198)
> => (counter=s@language:s@possible, value=200)
> => (counter=s@language:b@00000001:s@pt_BR, value=0)
> If I print out the byte value of the column names along with their MD5 sum I see:
> Name: [language, java.nio.HeapByteBuffer[pos=0 lim=4 cap=4], pt_BR]
> Byte array: ffffff8073086c616e67756167650ffffff80620400010ffffff80730570745f42520
> MD5: 2db353a9a72a0d7cb6cb277ac5125653
> Name: [language, possible]
> Byte array: ffffff8073086c616e67756167650ffffff807308706f737369626c650
> MD5: 82cad9b6a65c794e97cf1d4613e2e367
> Name: [language, java.nio.HeapByteBuffer[pos=0 lim=4 cap=4], pt_BR]
> Byte array: ffffff8073086c616e67756167650ffffff80620400010ffffff80730570745f42520
> MD5: 2db353a9a72a0d7cb6cb277ac5125653
> Unfortunately, I have been unable to duplicate this manually or via a generic test script
and our QA department can only duplicate ~25% of the time.

--
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