cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Brandon Williams (Updated) (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (CASSANDRA-3351) If node fails to join a ring it will stay in joining state indefinately
Date Thu, 20 Oct 2011 17:38:10 GMT

     [ https://issues.apache.org/jira/browse/CASSANDRA-3351?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Brandon Williams updated CASSANDRA-3351:
----------------------------------------

    Attachment: 3351.txt

Two problems introduced in CASSANDRA-2496: the fatclient logic was changed a bit too much,
and the isDeadState check was assuming fat clients had dead state, thus calling setHasToken
to effectively not mark  them as fat clients.
                
> If node fails to join a ring it will stay in joining state indefinately
> -----------------------------------------------------------------------
>
>                 Key: CASSANDRA-3351
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3351
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.8.6
>         Environment: xlarge ec2, Ubuntu 11.04 Natty, JNA
>            Reporter: Tuukka Luolamo
>            Assignee: Brandon Williams
>            Priority: Trivial
>              Labels: nodetool
>         Attachments: 3351.txt, cassandra.log.gz
>
>
> While attempting to add a new node to my ring something went wrong and I had to terminate
the node on ec2. After this the node keeps appearing in the ring command in "joining" state
and never goes away. Per driftx on the Cassandra channel if I do a whole cluster restart it
should go away, but since this is a production system this is not really possible. Additionally
if I could join a node with same IP again this should go away, but being on ec2 this is not
always easy. So not sure if this truly qualifies as a bug or more like a feature request,
but I feel there should be a way to remove a node in any state if I wish without joining a
node with same ip or doing a whole cluster restart.

--
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