I added an agenda items to talk about jenkins infra resource issues.

Is there anything we can do from an infra perspective to help with the
resource issues we see in the jenkins runs? Seems like any time the queue
gets over 200 jobs you can start to see issues. Restarting jobs seems like
a temporary workaround as it would just keep the problem going.

This seems to be a recurring issue. netvirt and genius have csit jobs to
gate patches and the job frequently time out after 6 hours. Our projects
rely heavily on these csit tests to gate patches and guarantee everything
is still working. Infra issues create a lot of issues during active
development. We also spend countless cycles tweaking and reworking tests to
squeak out tiny optimizations.

Around release times it is usually bad also. There is a flurry of activity,
everyone pushing patches and the job queues grow.

Thanks, Sam
_______________________________________________
Discuss mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/discuss

Reply via email to