#1192 don't let build_dispatcher blocked by failed initial createrepo
Closed: Fixed 3 years ago by praiskup. Opened 4 years ago by praiskup.

From time to time, some builds get stuck on waiting for createrepo to finish (build dispatcher doesn't process them), but if createrepo() it never continues (till I hit regenerate button).

Instead of waiting, it would be ideal to rather generate the repository (intiated by build_dispatcher)


Metadata Update from @msuchy:
- Issue assigned to thrnciar

4 years ago

Reopening, the situation still happens but the build leads to failure.

Metadata Update from @praiskup:
- Issue status updated to: Open (was: Closed)

3 years ago

Metadata Update from @praiskup:
- Assignee reset
- Issue untagged with: release-blocker

3 years ago

Seems like that the "create repo" actions are getting lost when new
chroots are enabled. Per irc report from @abbra:

<ab> praiskup: https://copr.fedorainfracloud.org/coprs/abbra/gc-wip/builds/
<ab> praiskup: last few builds should have succeeded but they aren't and no logs available
<ab> I started to regenerate repositories
<ab> I didn't have F32/Rawhide in past
<praiskup> ab yes, that would be #1192
<praiskup> there are no metadata: https://download.copr.fedorainfracloud.org/results/abbra/gc-wip/fedora-rawhide-x86_64/
<ab> because I never had it in past, only f31
<ab> then I changed settings to provide f32/rawhide
<ab> so I guess you need a logic that verifies there is metadata on the target change
<ab> and runs it otherwise

I temporarily did today:

--- /usr/bin/copr-repo  2020-02-05 09:48:40.000000000 +0000
+++ /usr/bin/copr-repo  2020-05-04 09:26:58.013108110 +0000
@@ -187,7 +187,10 @@

     try:
         if createrepo_run_needed:
-            run_cmd(createrepo_cmd, opts)
+            if '@python/python3.9' not in opts.directory:
+                run_cmd(createrepo_cmd, opts)
+            else:
+                opts.log.info("createrepo skipped for blacklisted projects")
         else:
             opts.log.info("createrepo_c run is not actually needed, "
                           "skipping command: %s",

Btw., on a weekly basis I need to advice poeple to re-generate the repodata
in projects which obviously have failed the initial createrepo request. Last
time:

<pvalena> hello, any idea why fedora-32-x86_64 doesn't work/build for me?
<pvalena> https://copr.fedorainfracloud.org/coprs/pvalena/rubygems/build/1401153/
<pvalena> https://copr.fedorainfracloud.org/coprs/pvalena/vagrant/build/1401098/

This issue is much less likely to happen now, the initial createrepo action is "prioritized"
so - even when there's huge backlog - it should be processed quickly.

I'm waiting a month or so, but this has potential to be closed.

We don't seem to face this issue anymore.

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

3 years ago

Login to comment on this ticket.

Metadata
Related Pull Requests
  • #1200 Closed 2 years ago
  • #1197 Merged 4 years ago