You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Till Rohrmann (JIRA)" <ji...@apache.org> on 2019/07/25 07:36:00 UTC
[jira] [Assigned] (FLINK-13407) Fix thread visibility of checked
SlotManager.failUnfulfillableRequest in StandaloneResourceManagerTest
[ https://issues.apache.org/jira/browse/FLINK-13407?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Till Rohrmann reassigned FLINK-13407:
-------------------------------------
Assignee: Andrey Zagrebin
> Fix thread visibility of checked SlotManager.failUnfulfillableRequest in StandaloneResourceManagerTest
> ------------------------------------------------------------------------------------------------------
>
> Key: FLINK-13407
> URL: https://issues.apache.org/jira/browse/FLINK-13407
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Coordination
> Affects Versions: 1.9.0
> Reporter: Andrey Zagrebin
> Assignee: Andrey Zagrebin
> Priority: Blocker
> Fix For: 1.9.0, 1.10.0
>
>
> One of the potential reasons of the test instability, described in FLINK-13242, is that the test checks _failUnfulfillableRequest_ directly in its thread but it is set in the main RPC thread of _StandaloneResourceManager#initialize_. It can lead to a potential problem with the visibility of _SlotManager.failUnfulfillableRequest_ in _StandaloneResourceManagerTest#testStartupPeriod_.
> The idea for the fix is to read _SlotManager.failUnfulfillableRequest_ in the main thread of _StandaloneResourceManager_, get future result and check in the _StandaloneResourceManagerTest.assertHappensUntil_.
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)