cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tyler Hobbs (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7836) Data loss after moving tokens
Date Tue, 26 Aug 2014 22:19:58 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7836?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14111454#comment-14111454
] 

Tyler Hobbs commented on CASSANDRA-7836:
----------------------------------------

It looks like this was introduced by CASSANDRA-3668.

{{StreamPlan.isEmpty()}} should be returning {{!coordinator.hasActiveSessions()}} instead
of {{coordinator.hasActiveSessions()}}.  This is currently only used when moving.

> Data loss after moving tokens
> -----------------------------
>
>                 Key: CASSANDRA-7836
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7836
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>         Environment: OSX and Ubuntu 14.04
>            Reporter: Philip Thompson
>            Assignee: Tyler Hobbs
>            Priority: Critical
>             Fix For: 2.1.0
>
>         Attachments: topology_test.py
>
>
> The dtest topology_test:TestTopology.movement_test is failing on 2.1.0 and 2.1-HEAD.
The test, which has been attached, goes through the following workflow:
> - Create an unbalanced three node cluster without vnodes
> - Create a keyspace with RF=1
> - Load data into the cluster
> - Flush the cluster with nodetool flush
> - Move the tokens around to balance the cluster, using nodetool move
> - Run nodetool cleanup
> At this point, the test checks that all 10,000 of the originally inserted rows are still
there. In 2.0.10 that is true, however in 2.1, only ~4000 rows exist.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message