[
https://issues.apache.org/jira/browse/YETUS-547?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16173418#comment-16173418
]
Allen Wittenauer commented on YETUS-547:
----------------------------------------
This kind of goes hand-in-hand with YETUS-504 . If the auth to download from a
bug system is only available outside of the container, then the download has to
happen before the Docker launch.
> Race condition when executing docker mode and an issue tracker gets updates
> ---------------------------------------------------------------------------
>
> Key: YETUS-547
> URL: https://issues.apache.org/jira/browse/YETUS-547
> Project: Yetus
> Issue Type: Bug
> Components: Test Patch
> Affects Versions: 0.5.0
> Reporter: Sean Busbey
> Priority: Minor
>
> Over in hbase we had a race condition where test patch's docker mode saw two
> different patches attached to an issue. The first one was used to pick the
> docker image and then on reexec we checked the issue tracker for a patch for
> a second time and got a different patch. The two patches were for different
> branches with different JVM requirements, which gave spurious failures.
> We should pass the patch we already downloaded in to the docker mode
> reexecution.
> Workaround: wait until a given run has gotten past the second patch finding
> before updating your issue tracker with a new patch.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
