If there's no record in PDC (it may happen), there are several retries. If it doesn't succeed, after the final retry, there's a failed handler for the failed module. At the end of the handler, there are component builds cancellations. But if there's no build.koji_tag (due to missing PDC record), it finally raises as follows:
failed
build.koji_tag
scheduler_1 | 2016-12-14 11:40:47,973 - module_build_service - ERROR - Failed while handling u'2016-c7ee9a26-7fc7-42b3-861d-1c73e660674f' scheduler_1 | Traceback (most recent call last): scheduler_1 | File "/tmp/module_build_service/module_build_service/scheduler/main.py", line 141, in run scheduler_1 | self.process_message(session, msg) scheduler_1 | File "/tmp/module_build_service/module_build_service/scheduler/main.py", line 171, in process_message scheduler_1 | further_work = handler(conf, session, msg) or [] scheduler_1 | File "/tmp/module_build_service/module_build_service/scheduler/handlers/modules.py", line 86, in failed scheduler_1 | builder.buildroot_connect(groups) scheduler_1 | File "/tmp/module_build_service/module_build_service/builder.py", line 445, in buildroot_connect scheduler_1 | self.tag_name, self.arches, perm="admin") # the main tag needs arches so pungi can dump it scheduler_1 | File "/tmp/module_build_service/module_build_service/builder.py", line 689, in _koji_create_tag scheduler_1 | taginfo = self.koji_session.getTag(tag_name) scheduler_1 | File "/usr/lib/python2.7/site-packages/koji/__init__.py", line 1943, in __call__ scheduler_1 | return self.__func(self.__name, args, opts) scheduler_1 | File "/usr/lib/python2.7/site-packages/koji/__init__.py", line 2379, in _callMethod scheduler_1 | raise err scheduler_1 | GenericError: invalid type for tagInfo: <type 'NoneType'>
@fivaldi changed the status to Closed
Closed
Fixed in PR#265.
Commit e0240ede fixes this issue
This issue has been migrated to Fedora Forge: https://forge.fedoraproject.org/koji/mbs/issues/264
Please continue any further discussion there.