pungi.global.log
[IMAGE_BUILD ] [INFO ] [FAIL] Image build (variant AtomicHost, arch *, subvariant AtomicHost) failed, but going on anyway. [IMAGE_BUILD ] [INFO ] ImageBuild task failed: 28602164. See /mnt/koji/compose/updates/Fedora-28-updates-testing-20180725.0/logs/aarch64-ppc64le-x86_64/imagebuild-AtomicHost-AtomicHost-qcow2-raw-xz.aarch64-ppc64le-x86_64.log for more details.
[IMAGE_BUILD ] [INFO ] [FAIL] Image build (variant AtomicHost, arch *, subvariant AtomicHost) failed, but going on anyway. [IMAGE_BUILD ] [INFO ] ImageBuild task failed: 28602165. See /mnt/koji/compose/updates/Fedora-28-updates-testing-20180725.0/logs/x86_64/imagebuild-AtomicHost-AtomicHost-vagrant-libvirt-vagrant-virtualbox.x86_64.log for more details.
@lsedlar - can you help me understand the GenericError: Build already in progress (task 28602158) issue. It thought it was because updates and updates-testing clashing with N-V-R in koji, but it doesn't happen often enough I don't think for it to be exactly that
GenericError: Build already in progress (task 28602158)
N-V-R
That error really means that a build started that would create an NVR that is already in progress. Indeed there's a successful build with that version: https://koji.fedoraproject.org/koji/buildinfo?buildID=1130719 (for task 28602165). The task that created the build started about 15 seconds before the failed one.
I'm really not sure how this could have happened. It's as if the command was ran twice, but I don't see any way how pungi could do that.
it's happened a few other times - anything we can do to try to track it down?
an attempt at a fix: https://lists.fedoraproject.org/archives/list/infrastructure@lists.fedoraproject.org/thread/MXEVNGFJHMJDXYCBMSCOSNJXB2MWVGFE/