flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ufuk Celebi (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (FLINK-1660) MultiTriggerPolicyTest timeout during test
Date Sat, 11 Apr 2015 09:18:12 GMT

    [ https://issues.apache.org/jira/browse/FLINK-1660?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14490881#comment-14490881
] 

Ufuk Celebi edited comment on FLINK-1660 at 4/11/15 9:18 AM:
-------------------------------------------------------------

Seems to be a thread-safety issue.

Each thread running ActiveTriggerWithRunnable calls the MyCallbackClass callback, which adds
an element to the linked list of received elements. This list is not thread-safe. That's why
the main thread running the test method sometimes waits indefinitely to receive all elements.

PS: This is an issue of the test not the tested component from what I've seen. I'll change
this and push the fix.


was (Author: uce):
Seems to be a thread-safety issue.

Each thread running ActiveTriggerWithRunnable calls the MyCallbackClass callback, which adds
an element to the linked list of received elements. This list is not thread-safe. That's why
the main thread running the test method sometimes waits indefinitely to receive all elements.

> MultiTriggerPolicyTest timeout during test
> ------------------------------------------
>
>                 Key: FLINK-1660
>                 URL: https://issues.apache.org/jira/browse/FLINK-1660
>             Project: Flink
>          Issue Type: Bug
>          Components: Streaming
>            Reporter: Ufuk Celebi
>            Priority: Minor
>             Fix For: 0.9
>
>
> Running tests through Maven locally, I encountered a timeout in 
> MultiTriggerPolicyTest#testActiveTriggerRunnables
> saying that it timed out after 10s.
> I've looked into the code and saw that it is almost totally unrelated to the runtime.
I therefore suspect that it was due to a system hickup. Could someone please confirm this
and maybe increase the timeout?
> In general, we should consider setting these timeouts rather high (due to Travis etc.)



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message