[
https://issues.apache.org/jira/browse/STORM-1934?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15354878#comment-15354878
]
ASF GitHub Bot commented on STORM-1934:
---------------------------------------
Github user satishd commented on a diff in the pull request:
https://github.com/apache/storm/pull/1528#discussion_r68914602
--- Diff: storm-core/src/clj/org/apache/storm/daemon/supervisor.clj ---
@@ -592,24 +577,13 @@
(doseq [p (set/difference (set (keys existing-assignment))
(set (keys new-assignment)))]
(.killedWorker isupervisor (int p)))
- (kill-existing-workers-with-change-in-components supervisor
existing-assignment new-assignment)
(.assigned isupervisor (keys new-assignment))
--- End diff --
`StandaloneSupervisor` implementation does nothing. What is the need of
this API?
> Race condition between sync-supervisor and sync-processes raises several
> strange issues
> ---------------------------------------------------------------------------------------
>
> Key: STORM-1934
> URL: https://issues.apache.org/jira/browse/STORM-1934
> Project: Apache Storm
> Issue Type: Bug
> Components: storm-core
> Affects Versions: 1.0.0, 2.0.0, 1.0.1
> Reporter: Jungtaek Lim
> Assignee: Jungtaek Lim
> Priority: Critical
>
> There're some strange issues including STORM-1933 and others (which I will
> file an issue soon) which are related to race condition in supervisor.
> As I mentioned to STORM-1933, basically sync-supervisor relies on zk
> assignment, and sync-processes relies on local assignment and local workers
> directory, but in fact sync-supervisor also access local state and take some
> actions which affects sync-processes. And also Satish left the comment to
> STORM-1933 describing other issue related to race condition and idea to fix
> this which is same page on me.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
