Closed alanshaw closed 6 years ago
Jobs shouldn't timeout on waiting for workers, where does it happen?
We could also try implementing dynamic worker scaling, trading lack of queues for slightly longer job startup (likely using https://wiki.jenkins.io/display/JENKINS/JClouds+Plugin)
@magik6k I can't really send you a link to an example as the worker becomes available and the tests start but then get killed after the run takes more than an hour.
Ok, well this is probably an instance where this happened: https://ci.ipfs.team/blue/organizations/jenkins/IPFS%2Fjs-ipfs/detail/PR-1576/1/pipeline/16
Is there a way to enforce the time limit after all the jobs in the pipeline have been assigned a worker? When CI is busy it can take ages to get a worker and so the run fails. It wastes resources and my time having to retstart it again.