[
https://issues.apache.org/jira/browse/STORM-375?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14213939#comment-14213939
]
ASF GitHub Bot commented on STORM-375:
--------------------------------------
Github user xiaokang commented on the pull request:
https://github.com/apache/storm/pull/167#issuecomment-63222238
In our production storm cluster with 600 nodes, ZK read bandwidth had been
reduced by 75% by applying this PR and simple fix for the "swap!" mistake I
mentioned above.
However the mistake of using "swap!" instead of "reset!" in (swap!
(:assignment-versions supervisor) versions) is still in master branch.
@knusbaum & @nathanmarz , any idea?
> Smarter downloading of assignments by supervisors and workers
> -------------------------------------------------------------
>
> Key: STORM-375
> URL: https://issues.apache.org/jira/browse/STORM-375
> Project: Apache Storm
> Issue Type: Improvement
> Reporter: Kyle Nusbaum
> Assignee: Kyle Nusbaum
> Fix For: 0.9.3
>
>
> Periodically the supervisors and workers each download their assignments from
> zookeeper, and the traffic begins to saturate the network and hammer
> zookeeper as we scale. Having these daemons only grab the assignments when
> the zookeeper path's version number changes reduces this traffic
> considerably, along with STORM-376.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
