You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@geode.apache.org by "ASF subversion and git services (Jira)" <ji...@apache.org> on 2021/09/16 19:44:00 UTC
[jira] [Commented] (GEODE-9495) remove thread sleep from
PubSubNativeRedisAcceptanceTest class cleanup
[ https://issues.apache.org/jira/browse/GEODE-9495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17416309#comment-17416309 ]
ASF subversion and git services commented on GEODE-9495:
--------------------------------------------------------
Commit ed902f6404a1c2b06b980973a7a9d74f045fee8b in geode's branch refs/heads/develop from Ray Ingles
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=ed902f6 ]
GEODE-9495: Limit Thread Sleep In PubSubNativeRedisAcceptanceTest (#6866)
Dynamically adjusts the wait time at the end of the test based on OS values (to allow many sockets to close before subsequent tests).
Co-authored-by: Ray Ingles <ri...@vmware.com>
> remove thread sleep from PubSubNativeRedisAcceptanceTest class cleanup
> ----------------------------------------------------------------------
>
> Key: GEODE-9495
> URL: https://issues.apache.org/jira/browse/GEODE-9495
> Project: Geode
> Issue Type: Improvement
> Components: redis
> Affects Versions: 1.15.0
> Reporter: Hale Bales
> Assignee: Ray Ingles
> Priority: Major
> Labels: pull-request-available
>
> GEODE-9338 includes the addition of a Thread.sleep(240000) in the class cleanup. This wait is necessary to allow the sockets used in this class to exit the TIME_WAIT state. Out of Mac, Linux, and Windows, Windows has the longest default TIME_WAIT, which is 240 seconds. So currently at the end of the PubSubNativeRedisAcceptanceTest tests, we wait for 240 seconds.
> Another solution was proposed that involved parsing netstat outputs to find out when enough sockets have been freed, but the complexity didn't buy us much there.
> A potential other solution would be to reduce the TIME_WAIT period when running the tests in CI.
> If we don't wait for the TIME_WAIT period to be up then other tests being run after it will get bind exceptions.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)