On Thu, 2019-11-14 at 12:44 -0800, Adam Williamson wrote: > On Thu, 2019-11-14 at 10:34 -0800, Kevin Fenzi wrote: > > On Thu, Nov 14, 2019 at 08:13:52AM -0800, Adam Williamson wrote: > > > On Wed, 2019-11-13 at 14:32 -0800, Kevin Fenzi wrote: > > > > autocloud*: f27, but should die when f29 goes eol (it's only used to > > > > test atomic-host). > > > > > > No, it isn't. It also tests the regular cloud images: > > > > Sigh. :( > > > https://apps.fedoraproject.org/autocloud/jobs/ > > > > > > the tests with family 'Base' are tests of non-Atomic images. The > > > regular Cloud base images are still release-blocking until whenever we > > > decide to switch over to CoreOS. At some point this testing was meant > > > > Well, or decide to just make the Cloud base image not blocking. > > > > > to move to Taskotron, then there was some talking of moving it to > > > Fedora CI, but I don't know if anyone has actually done any work on > > > either of those things. AFAIK we have no other system currently for > > > doing automated testing on the Cloud base images. > > > > > > We could probably run the tests in openQA without too much trouble if > > > desired (though I can't guarantee it as I haven't looked into it in any > > > detail), but we should probably do *something* here. > > > > yeah, keeping autocloud is really not an option. > > It's running on a eol OS. > > > > We have no upstream developers anymore. > > > > It takes up 2 large hardware boxes we can't use for anything else that > > are idle most of the time. > > > > So, yes, we should come up with something here. ;( > > welp, if no-one else has a plan, I can take a look at migrating the > tests it runs into openQA maybe next week. (That might also mean we'd > pay more attention to the results - the autocloud tests haven't passed > on Rawhide for like two months, but I didn't know that until I looked > this morning...) Well that turned out to be not too hard! https://openqa.stg.fedoraproject.org/tests/674207 So I briefly looked at whether I wanted to take over autocloud but there's a lot of it, and it looks to me like ultimately all it really *does* is boot the image and then do this: https://infrastructure.fedoraproject.org/cgit/ansible.git/tree/roles/autocloud/backend/files/fedora.txt where ## means "this shouldn't fail the overall test if it fails", @@ means "this is expected to fail" and POLL means "wait till the system reboots" (per https://github.com/kushaldas/tunir/blob/master/docs/usage.rst). So I just taught openQA to be able to boot Cloud_Base qcow2 images and do basically the same: https://openqa.stg.fedoraproject.org/tests/674207/modules/autocloud/steps/1/src I'll get this cleaned up in various ways and deployed to production over today and tomorrow. It doesn't replicate *all* of autocloud - most notably for now it'll only test the Cloud_Base qcow2 images, not the vagrant images which autocloud also tested - but I think that's good enough for now. I'm not sure if it makes more sense to try and make openQA mimic the fedmsgs that autocloud currently publishes and the resultsdb results it reports, or just find anything that uses either of those and make it work with the openQA messages and results instead, I'll look into that. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net _______________________________________________ infrastructure mailing list -- infrastructure@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to infrastructure-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/infrastructure@xxxxxxxxxxxxxxxxxxxxxxx
