cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Patricio Echague (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-3082) Operation with CL=EACH_QUORUM doesn't succeed when a replica is down (RF=3)
Date Fri, 26 Aug 2011 19:19:29 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-3082?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13091966#comment-13091966
] 

Patricio Echague commented on CASSANDRA-3082:
---------------------------------------------

It passed a manual test using CCM with 6 nodes.

DC1:3
DC2:3

Originally, with all replicas up and running, the insert with CL=EACH_QUORUM failed.

The patch addresses that issue.

> Operation with CL=EACH_QUORUM doesn't succeed when a replica is down (RF=3)
> ---------------------------------------------------------------------------
>
>                 Key: CASSANDRA-3082
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3082
>             Project: Cassandra
>          Issue Type: Bug
>    Affects Versions: 0.7.0, 0.8.0
>            Reporter: Patricio Echague
>            Assignee: Patricio Echague
>              Labels: consistency
>             Fix For: 0.7.9, 0.8.5, 1.0
>
>         Attachments: CASSANDRA-0.7-3084.txt
>
>
> {code}  DatacenterSyncWriteResponseHandler#assureSufficientLiveNodes()
>      ...
>      ...
>         // Throw exception if any of the DC doesn't have livenodes to accept write.
>         for (String dc: strategy.getDatacenters())
>         {
>         	if (dcEndpoints.get(dc).get() != responses.get(dc).get())
>                 throw new UnavailableException();
>         }
> {code}
> should be:
>  
> {code}
>       if (dcEndpoints.get(dc).get() < responses.get(dc).get())
> {code}

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message