[jira] [Updated] (CASSANDRA-2557) StorageProxy sends same message multiple times

["Shotaro Kamio (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/CASSANDRA-2557?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Shotaro Kamio updated CASSANDRA-2557:
-------------------------------------

    Attachment: StorageProxy.java.patch

The problem occurs because the sendMessages() is within for-loop of StorageProxy. A patch to fix that is attached.

> StorageProxy sends same message multiple times
> ----------------------------------------------
>
>                 Key: CASSANDRA-2557
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2557
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.8 beta 1
>         Environment: linux
>            Reporter: Shotaro Kamio
>         Attachments: StorageProxy.java.patch
>
>
> A cassandra node gets multiple mutation messages (in number of times of replication factor at maximum) for an insert. It may cause high load on the node. The mutation should be only once for each insert.
> This bug is visible via MutationStage count in nodetool tpstats.
> For instance, if you have 6 node cluster (initial keys are 31, 32, 33, 34, 35 and 36) with replication factor = 4 and a single data (for example, key='2') is inserted, MutationStage count will be as follows:
> node 1: MutationStage 0 0 4
> node 2: MutationStage 0 0 3
> node 3: MutationStage 0 0 2
> node 4: MutationStage 0 0 1
> node 5: MutationStage 0 0 0
> node 6: MutationStage 0 0 0
> As you can see, the counts are different in each node.

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