www-infrastructure-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Wes McKinney (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (INFRA-11293) More Travis CI capacity for ASF projects
Date Tue, 01 Mar 2016 18:16:18 GMT

    [ https://issues.apache.org/jira/browse/INFRA-11293?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15174133#comment-15174133
] 

Wes McKinney commented on INFRA-11293:
--------------------------------------

Thanks. Part of the problem is that commits to trunk also trigger builds. So if you have a
commit flow where you merge a PR as soon as the build passes, you may be immediately triggering
another 2 hour build on 5 nodes. Right this moment, for example, there are 2 active builds
on master: https://travis-ci.org/apache/flink/branches and 2 builds on PRs https://travis-ci.org/apache/flink/pull_requests,
so Flink is occupying 20 nodes near as I can tell. 

> More Travis CI capacity for ASF projects
> ----------------------------------------
>
>                 Key: INFRA-11293
>                 URL: https://issues.apache.org/jira/browse/INFRA-11293
>             Project: Infrastructure
>          Issue Type: Wish
>          Components: Github
>            Reporter: Wes McKinney
>
> As reported in FLINK-3452, high Travis utilization by a small subset of ASF projects
on GitHub has challenged developer productivity on smaller projects with faster builds.
> Here is a total duration in seconds for each of the apache's active projects for the
last 180 days (provided to me by Travis CI): https://gist.githubusercontent.com/drogus/d01a02aa9e0f7c380bde/raw/05be62eb1dd76085dd6cc9740e496ac2a111bc33/duration-by-project.txt
> I would suggest increasing our capacity beyond the 30 concurrent workers currently available
or exploring procuring dedicated build slaves for projects with high overall utilization.



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

Mime
View raw message