Closed SethTisue closed 9 years ago
I've been thinking about this, and I like the idea of setting no timeout by default, and having job-specific timeouts for jobs where we expect them to sometimes not end.
I'll submit a PR to concretize my thoughts in a couple of minutes.
looking at https://scala-ci.typesafe.com/job/scala-2.11.x-validate-test/buildTimeTrend, since we've limited concurrency to 2, no build took more than 1h47min
Hm, the concurrency change is more recent, so I should've said, in the last 2 months, no build took 2hrs.
we should also really configure job concurrency at a category level:
hudson.plugins.throttleconcurrents.ThrottleJobProperty.xml:
<?xml version='1.0' encoding='UTF-8'?>
<hudson.plugins.throttleconcurrents.ThrottleJobProperty_-DescriptorImpl plugin="throttle-concurrents@1.8.4">
<categories>
<hudson.plugins.throttleconcurrents.ThrottleJobProperty_-ThrottleCategory>
<maxConcurrentPerNode>2</maxConcurrentPerNode>
<maxConcurrentTotal>0</maxConcurrentTotal>
<categoryName>cpu-heavy</categoryName>
<nodeLabeledPairs/>
</hudson.plugins.throttleconcurrents.ThrottleJobProperty_-ThrottleCategory>
</categories>
</hudson.plugins.throttleconcurrents.ThrottleJobProperty_-DescriptorImpl
this is n/a anymore after #103 was merged and #102 was opened
see https://github.com/scala/scala-jenkins-infra/pull/100 for details