You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Héctor Izquierdo (JIRA)" <ji...@apache.org> on 2011/07/06 16:15:17 UTC
[jira] [Created] (CASSANDRA-2865) During repair mark a node as
being repared, so no reads go to that node
During repair mark a node as being repared, so no reads go to that node
-----------------------------------------------------------------------
Key: CASSANDRA-2865
URL: https://issues.apache.org/jira/browse/CASSANDRA-2865
Project: Cassandra
Issue Type: Improvement
Components: Core
Affects Versions: 0.8.1
Reporter: Héctor Izquierdo
If a disk breaks and you lose a node data, when you bring it up again to do the repair, it will serve reads, and if clients are using CL.ONE, they will get bad data. Would it be possible to signal somehow that the node should not be trusted and reads should go to any other replica?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (CASSANDRA-2865) During repair mark a node as
being repared, so no reads go to that node
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-2865?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis resolved CASSANDRA-2865.
---------------------------------------
Resolution: Not A Problem
The right way to fix this is to removetoken the node w/ the bad disk, and rebootstrap it once it is repaired. See http://wiki.apache.org/cassandra/Operations
> During repair mark a node as being repared, so no reads go to that node
> -----------------------------------------------------------------------
>
> Key: CASSANDRA-2865
> URL: https://issues.apache.org/jira/browse/CASSANDRA-2865
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 0.8.1
> Reporter: Héctor Izquierdo
>
> If a disk breaks and you lose a node data, when you bring it up again to do the repair, it will serve reads, and if clients are using CL.ONE, they will get bad data. Would it be possible to signal somehow that the node should not be trusted and reads should go to any other replica?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-2865) During repair mark a node as
being repared, so no reads go to that node
Posted by "Héctor Izquierdo (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-2865?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13060654#comment-13060654 ]
Héctor Izquierdo commented on CASSANDRA-2865:
---------------------------------------------
You are absolutely right. But wouldn't it still be useful when you are repairing a node and it can take several hours to complete? Why don't treat it as bootstrap?
> During repair mark a node as being repared, so no reads go to that node
> -----------------------------------------------------------------------
>
> Key: CASSANDRA-2865
> URL: https://issues.apache.org/jira/browse/CASSANDRA-2865
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 0.8.1
> Reporter: Héctor Izquierdo
>
> If a disk breaks and you lose a node data, when you bring it up again to do the repair, it will serve reads, and if clients are using CL.ONE, they will get bad data. Would it be possible to signal somehow that the node should not be trusted and reads should go to any other replica?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira