Follow-up from "Add Pipeline Schedules that supersedes experimental Trigger Schedule"
The following discussion from !10853 (merged) should be addressed:
-
@ayufan started a discussion:
Should we have foreign keys for
owner_id
?
Do not update/delete: Banner broadcast message test data
Do not update/delete: Notification broadcast message test data
The following discussion from !10853 (merged) should be addressed:
@ayufan started a discussion:
Should we have foreign keys for
owner_id
?
Maybe we should also open an issue on supporting variables
@ayufan I wonder what the best way is to move forward with this, most obvious is to create a FK with nullify. However, this might break the pipelines triggered after the owner is deleted. So far its not impossible to nullify and set active to false.
We could skip the this situation in the worker, but we might also just accept this imperfect solution for now, as its quick and works most of the time.
Yes. I think that pipelines created from the schedule with nullified owner should be failed by default.
mentioned in commit c8928248
mentioned in merge request !11233 (merged)
mentioned in commit 2d147397
mentioned in commit a09b0d2b
mentioned in commit 02e156c2
closed via merge request !11233 (merged)
mentioned in commit nick.thomas/gitlab-ee@cc572c43
mentioned in commit test-nick/gitlab-ce@0919a0ed
mentioned in merge request !14785