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 2019/12/05 00:45:00 UTC
[jira] [Commented] (GEODE-7540) avoid overwhelming CI when many
commits happen on the same day
[ https://issues.apache.org/jira/browse/GEODE-7540?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16988344#comment-16988344 ]
ASF subversion and git services commented on GEODE-7540:
--------------------------------------------------------
Commit 3c524cb8cae164e70f519f3eed0ccf17f1914d87 in geode's branch refs/heads/develop from Owen Nichols
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=3c524cb ]
GEODE-7540: use max_in_flight to spread out peak load on CI (#4423)
* GEODE-7540: use max_in_flight to spread out peak load on CI when many commits come on the same day
* retries rarely help and mostly just burn up containers, reduce all retry counts to 2
> avoid overwhelming CI when many commits happen on the same day
> --------------------------------------------------------------
>
> Key: GEODE-7540
> URL: https://issues.apache.org/jira/browse/GEODE-7540
> Project: Geode
> Issue Type: Improvement
> Components: ci
> Reporter: Owen Nichols
> Priority: Major
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Instead of immediately launching all jobs at once, we could even out the load by setting lower max_in_flight settings for shorter jobs. I estimate we can cut peak load in half without builds taking any longer, for a scenario where 6 commits are in the pipeline at the same time.
> Reducing load on concourse is desirable because we tend to have more unexplained rsync failures, etc, when concourse is under heavy load.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)