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
{{ message }}
Repository navigation
OSX machines are de-provisioned during CI / PR runs leading to failures #34472
Because of the AzDO limitation where you can't rerun individual lanes until all of them are completed, this issue is particularly painful when trying to land a PR :(
Update from my end; chatted with AzDO folks in charge of the hosted macs today, the previous ETA of 7/15 to migrate the hosted agents seems to have not been accurate. I am politely asking for an updated time.
Yes, definitely still a thing (and why the issue is still open). I checked this instance out, and it's definitely on the v2 pool so the migration is not something we can hope for to help us here any more. I did share this specific instance back with the MMS team though they are tracking them en masse as well via telemetry.
Activity
Dotnet-GitSync-Bot commentedon Apr 2, 2020
I couldn't figure out the best area label to add to this issue. Please help me learn by adding exactly one area label.
jaredpar commentedon Apr 2, 2020
FYI this is not runtime specific. Hitting other repos as well
Evaluated 500 builds
Impacted 5 bulids
Impacted 6 jobs
dougbu commentedon Apr 21, 2020
This is also tracked at dotnet/core-eng#9448. Correct?
MattGal commentedon Apr 21, 2020
Correct.
AraHaan commentedon May 19, 2020
aww that link 404d for me.
MattGal commentedon May 19, 2020
Private repo; I can publicly tell you however that the problem continues to be investigated and seems to be at the MacOS-VM-provider level.
AraHaan commentedon May 19, 2020
Seems on my pr 36694 that this might be a cause when I read this: https://helix.dot.net/api/2019-06-17/jobs/b0b1c468-c4c5-4f2d-a2a6-3c831ed6ef20/workitems/System.Net.Sockets.Tests/console
MattGal commentedon May 19, 2020
No that log is from a plain old crash on a Helix machine. This issue is specific to hosted Azure Devops build macs, which are in no way connected.
CoffeeFlux commentedon May 22, 2020
Because of the AzDO limitation where you can't rerun individual lanes until all of them are completed, this issue is particularly painful when trying to land a PR :(
myblindy commentedon Jul 13, 2020
Any idea why my https://dev.azure.com/dnceng/public/_build/results?buildId=726102 build was included in this? It didn't fail any tests or build steps as far as I can tell.
MattGal commentedon Jul 13, 2020
It does on Attempt #1 :

So, if you didn't click retry, some nice person did for you :)
MattGal commentedon Jul 13, 2020
Update from my end; chatted with AzDO folks in charge of the hosted macs today, the previous ETA of 7/15 to migrate the hosted agents seems to have not been accurate. I am politely asking for an updated time.
directhex commentedon Sep 11, 2020
Seemingly still a thing https://dev.azure.com/dnceng/internal/_build/results?buildId=810664&view=logs&j=04b1677b-aa91-548a-e789-89c3d998d0f7
MattGal commentedon Sep 11, 2020
Yes, definitely still a thing (and why the issue is still open). I checked this instance out, and it's definitely on the v2 pool so the migration is not something we can hope for to help us here any more. I did share this specific instance back with the MMS team though they are tracking them en masse as well via telemetry.
ViktorHofer commentedon Oct 1, 2020
@MattGal we are now live updating this issues top post. Maybe that helps?
ViktorHofer commentedon Feb 12, 2021
Got fixed by AzDO. Closing.