You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Mgmt-plane-only PR triggers data-plane CI, for example #11625 , so if there's error in data-plane, the pr is blocked (check-enforcer demands all CI to pass).
Question
I think in general the failure of data-plane CI should not block mgmt-plane PRs, but I want to here your voice.
Proposed solution
The reviewer tells check enforcer it's OK to ignore the failed data-plane checks
Update data-plane CI to exclude mgmt-plane projects
We could simply let them continue to trigger and if they are failing for other reasons ignore the failures and proceed with check-in.
We could work to ensure both mgmt and data-plane stay clean and always passing.
I was chatting with @allenjzhang about the track 2 mgmt libraries and for those I would like to go with option 3. However for the existing track 1 mgmt libraries perhaps we can go with option 1. It shouldn't be too hard to alter that script to adjust all these.
This issue is fairly stale by now and the T1 packages are deprecated with T2 reaching GA. I'm going to close this out and we can address any specific cases on an as-needed basis.
What I see
Mgmt-plane-only PR triggers data-plane CI, for example #11625 , so if there's error in data-plane, the pr is blocked (check-enforcer demands all CI to pass).
Question
I think in general the failure of data-plane CI should not block mgmt-plane PRs, but I want to here your voice.
Proposed solution
cc @weshaggard @mitchdenny @erich-wang @chidozieononiwu
The text was updated successfully, but these errors were encountered: