Ref Bot does not take correct project branch into account

mentioned in merge request lhcb-core/LHCbNightlyConf!1213 (merged)

I have removed Boole from 2024-patches with lhcb-core/LHCbNightlyConf!1213 (merged). Can be added back once the issue here is resolved.

I started some refactoring some time ago to improve the way slots and branches are deduced, but it's in my spare time and it was not too urgent.

I'll try to work on it a bit more, unless RTA provides a maintainer for the project.

mentioned in merge request !48 (merged)

For the record, the problem in the refbot job mentioned in this MR doesn't seem related to Boole:

WARNING [update_references.py::285] Error during git operations, no reference update MR created for Boole!
INFO    [update_references.py::842] Successfully updated references for: ['Moore', 'DaVinci', 'Alignment']
error: RPC failed; HTTP 500 curl 22 The requested URL returned error: 500
send-pack: unexpected disconnect while reading sideband packet
fatal: the remote end hung up unexpectedly
Everything up-to-date
ERROR   [main.py::108] Caught Exception, sending reply to the trigger MR that something failed!
Command 'git -C Moore push --force -o ci.skip origin HEAD' returned non-zero exit status 1.
Traceback (most recent call last):
  File "/builds/lhcb-rta/reference-update-bot/reference_update_bot/main.py", line 70, in main
    sp.check_call(f"git -C {proj} push --force -o ci.skip origin HEAD", shell=True)
  File "/cvmfs/lhcb.cern.ch/lib/var/lib/LbEnv/3149/stable/linux-64/lib/python3.12/subprocess.py", line 413, in check_call
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command 'git -C Moore push --force -o ci.skip origin HEAD' returned non-zero exit status 1.

From the look of it, unknown branches are ignored and the failure was a transient problem trying to push to Moore repository.

Hmm, yes I missed that. I was perhaps too focussed on

fatal: couldn't find remote ref 2024-patches

during the Boole part..

@clemenci One reason I slightly question the idea this is a 'transient' issue is I have had situations where one MR would, repeatedly fail, multiple times in a row, but at the same time other MRs would be OK. I specifically at one point ran the ref bot on a few ref changes MRs in quick succession to test this. If it where a pure transient issue, I don't see how it would know to fail for one MR, but work OK for another, at essentially the same time ?

I guess we should maybe now just wait and see if it happens again, with your recent changes and also without Boole in 2024-patches. If it does, we can then poke at it a bit.

@msaur FYI as incoming maintainer.

Ref Bot does not take correct project branch into account

Designs

Child items 0

Activity