This message was deleted.
# atlantis-communitys
Slackbot
10/20/2023, 12:52 PMThis message was deleted.
😞 2
b
Bruno Schaatsbergen
10/20/2023, 12:55 PMSomeone created a PR (PR 1):
A plan was ran, and a lock was acquired
Meanwhile, another PR (PR 2) was raised by another engineer - when planning PR 1 again, it suddenly said: you're PR is locked by PR2
Even though PR 1 should've had the lock. There's also no sign of unlocking PR 1 (see below screenshot)
Bruno Schaatsbergen
10/20/2023, 12:59 PMThis is PR 2, which was raised after PR 1, and somehow magically removed the lock from PR 1 and started planning
Bruno Schaatsbergen
10/20/2023, 1:04 PMDo we delete the plan/lock if a plan fails now?
It seems like the lock and plan were removed after a failed plan, PR 2 then got raised and acquired a lock and started planning - causing PR 1 to go "on hold"
Bruno Schaatsbergen
10/20/2023, 1:13 PMThis could be some super odd edge case though, not quite sure what caused it
s
Simon Heather
10/20/2023, 1:18 PMLooks like you pushed two commits in quick succession to PR 2 which is why you got the plan error.
b
Bruno Schaatsbergen
10/20/2023, 1:19 PMI would've expected a lock on PR 2, as PR 1 got planned first - and never got unlocked by anyone
s
Simon Heather
10/20/2023, 1:20 PMCan't comment because I don't know the relative timings of each one.
b
Bruno Schaatsbergen
10/20/2023, 1:20 PMPR 2 got created several minutes after PR 1
Bruno Schaatsbergen
10/20/2023, 1:20 PMYou can see that in the screenshots
s
Simon Heather
10/20/2023, 1:21 PMBut you have pushed several commits to PR1. The first one was a failed plan so there would have been no lock.
b
Bruno Schaatsbergen
10/20/2023, 1:21 PMThe first one succeeded actually, but I couldn't capture that in a single screenshot
Bruno Schaatsbergen
10/20/2023, 1:22 PMPR 1 got made, a plan got ran, a commit got pushed, another plan got ran which error'd, meanwhile PR 2 got made after the first succesful plan of PR 1, and it then somehow got skewd
Bruno Schaatsbergen
10/20/2023, 1:22 PMThis is probably some weird edge case, but it seems reproducible on my end
s
Simon Heather
10/20/2023, 1:23 PMYou know you can see current locks via the Atlantis web interface.
Simon Heather
10/20/2023, 1:23 PMSo if you can reproduce it, raise a GitHub issue with the details.
b
Bruno Schaatsbergen
10/20/2023, 1:24 PMSure 🙂
s
Simon Heather
10/20/2023, 1:29 PMJust a reminder, Atlantis is maintained by volunteers and the community. You are more than welcome to look at the code and raise PRs to resolve any issues you find.
b
Bruno Schaatsbergen
10/20/2023, 1:29 PMI'm well aware of that, but thanks
p
PePe Amengual
10/20/2023, 3:54 PMtry to run another PR that have a successful plan/apply and see if that works as expected, it could be an edge case
d
Dylan Page
10/23/2023, 1:47 PMThis is exactly what I'm trying to resolve with the ADR
✅ 1
🙌 1