On Thu, Jun 18, 2020, at 11:55 AM, Sean McGinnis wrote: > > >> Interesting. The tag was pushed but upload-git-mirror role failed: > >> https://22ce6cebe995c8915d66-868b7eb6ba44b32e3b412bc5acf60e73.ssl.cf5.rackcdn.com/de51e32e9ff5ffedb6b66e6fcbc4a2ec9d4d5ddc/release/openstack-upload-github-mirror/fd8c86f/job-output.txt > >> > >> Is this a known issue? > > Was this a re-enqueue? It failed because the tag already existed on the remote: > > > > 2020-06-18 18:07:39.509620 | localhost | ! [remote rejected] 15.6.0 -> 15.6.0 (cannot lock ref 'refs/tags/15.6.0': reference already exists) > > > > If the content on github looks up to date and correct then I don't think there is anything else to do here. > > I don't think this was a reenqueue. Do we have another process that > syncs things to GitHub? Maybe a race between the release pushing the > mirror update and something else sneaking in and syncing it first? > > Either way, looks like things are up to date on GitHub and nothing else > to do here. > > Sean There were two github mirror jobs for puppet-nova running at roughly the same time. One succeeded (and pushed the tag) and the other failed. SUCCESS: https://zuul.openstack.org/build/1f363ade2d8d4f418db86ae33d2c3363/log/job-output.txt#80 FAILURE: https://zuul.openstack.org/build/fd8c86f9b57947239bd0273424e1d4d5/log/job-output.txt#80 This looks like a race between the job for the 16.4.0 and 15.6.0 tags. We could update the job to check for this particular failure case and treat it as a success assuming something else has merged. We could serialize these jobs between different events on the same repo. We can also just ignore it for now since everything is happy :)