#10197 cancelled build of glpk present in rawhide systems used in CI
Closed: Invalid 2 years ago by vkadlcik. Opened 2 years ago by vkadlcik.

Describe what you would like us to do:


I created a PR to introduce gating for the glpk component:
https://src.fedoraproject.org/rpms/glpk/pull-request/1

CI triggered a test run:
https://osci-jenkins-1.ci.fedoraproject.org/job/fedora-ci/job/dist-git-pipeline/job/master/71331/testReport/(root)/tests/

From the report it's obvious the test system has glpk-5.0-3.fc36.x86_64 present which is a cancelled build:
https://koji.fedoraproject.org/koji/buildinfo?buildID=1825617

Please make sure cancelled builds don't leak into composes/images used in CI. Further, please update the images so they have a latest build of glpk. Thanks.

When do you need this to be done by? (YYYY/MM/DD)


2021/09/10 would be nice :-)


Metadata Update from @mohanboddu:
- Issue priority set to: Waiting on Assignee (was: Needs Review)
- Issue tagged with: medium-gain, medium-trouble, ops

2 years ago

When testing pull requests, CI scratch-builds whatever is in the source branch.

The cancelled build just happens to have the same NVR (but it was a proper build, not CI scratch build).

So there is nothing to do here, I think.

@msrb OK, thanks Michal for the explanation. Now I see why "koji download-build" sometimes didn't work for me - it was just a wrong tool. Is there another way to download the scratch RPMs being tested (including the SRPM)? Thanks!

@msrb OK, thanks Michal for the explanation. Now I see why "koji download-build" sometimes didn't work for me - it was just a wrong tool. Is there another way to download the scratch RPMs being tested (including the SRPM)? Thanks!

To answer my question: The scratch-built RPMs (including the SRPM) are available during CI testing in a local repo called "test-artifacts". The repo has the highest priority. The easiest way to get the SRPM is:
yumdownloader --source <pkg>
(As a backup method the following can be used:
cp /var/share/test-artifacts/<NVR>.src.rpm <destdir>
but that relies on internals how the "test-artifacts" repo is implemented.)

This issue can be closed as invalid: It came from my misunderstanding of CI's scratchbuilding and a coincidence...

Metadata Update from @vkadlcik:
- Issue close_status updated to: Invalid
- Issue status updated to: Closed (was: Open)

2 years ago

Login to comment on this ticket.

Metadata
Boards 1
ops Status: Done