[
https://issues.apache.org/jira/browse/STORM-739?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rick Kellogg updated STORM-739:
-------------------------------
Component/s: storm-core
> If a worker dies supervisor clears local topology jar and re-downloads from
> nimbus
> ----------------------------------------------------------------------------------
>
> Key: STORM-739
> URL: https://issues.apache.org/jira/browse/STORM-739
> Project: Apache Storm
> Issue Type: Bug
> Components: storm-core
> Reporter: Sriharsha Chintalapani
> Assignee: Sriharsha Chintalapani
>
> Currently if a worker dies supervisor clears the local state including the
> supervisor's local topology jar. This is causing supervisor to re-download
> the topology jar from nimbus which can bring down nimbus if a topology has
> exception causing it to kill worker jvms and this can get into a loop worker
> dies-> supervisor clears the state- > supervisor downloads jar from nimbus .
> This will increase the load on the nimbus and there is no need to clear the
> supervisor state.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
