[jira] [Updated] (STORM-260) Race condition in backtype.storm.utils.Time

["Adrian Petrescu (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/STORM-260?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Adrian Petrescu updated STORM-260:
----------------------------------

    Attachment: patch_commit_ae6a0eaf21e1.patch

Patch to fix this issue.

> Race condition in backtype.storm.utils.Time
> -------------------------------------------
>
>                 Key: STORM-260
>                 URL: https://issues.apache.org/jira/browse/STORM-260
>             Project: Apache Storm (Incubating)
>          Issue Type: Bug
>    Affects Versions: 0.9.2-incubating
>            Reporter: Adrian Petrescu
>            Priority: Minor
>              Labels: patch, test
>         Attachments: patch_commit_ae6a0eaf21e1.patch
>
>
> Some of my test runs were occasionally failing with a NullPointerException on [backtype.storm.utils.Time.java:64|https://github.com/apache/incubator-storm/blob/master/storm-core/src/jvm/backtype/storm/utils/Time.java#L64].
> After a bit of investigation, it seems there's a race condition here; if we disable simulating mode while a thread is currently sleeping, then when it wakes up it won't re-check if it's still in "simulating" mode, it'll try to remove the sleep time, and get the NPE.
> The attached patch is a fairly straightforward fix.



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