[Podman-monitor] Re: Cirrus-CI cron build failures on containers/podman

On 4/22/21 8:06 PM, Do Not Reply wrote: This one looks like a legit problem in 'Test Code Consistency'. However, I checked and recent builds on master have passed fine. I'm going to assume the problem was caught and fixed. Someone wiser than me might go and check to verify, just to be safe. Another example of failure I asked support about: "Failed to start an instance! Pod failed!" (In case you see it in a PR) Support replied to me after digging deeper. They found this is a problem caused by our container images being so huge. They're seeing getting timeouts from quay (on occasion) downloading them. Unfortunately the image-size issue is likely due to building them as a full-blown environment to match our VMs setup. This is needed since we compile podman/buildah/skopeo and need lots of tooling to test them :S I can attempt to implement some tricks (like cleaning dnf cache and $GOCACHE) but I'm not sure about removing any packages. If anyone can help spot extranious/legacy packages they're sure we don't need, please let me know! https://github.com/containers/automation_images/blob/master/cache_images/... and https://github.com/containers/automation_images/blob/master/cache_images/... -- Chris Evich (he/him), RHCA III Senior Quality Assurance Engineer If it ain't broke, yain't tryin' hard 'nough.