#12039 libva-2.21.0-2.fc40 doesn't appear in repo
Closed: Fixed with Explanation by kevin. Opened by kwizart.

  • Describe the issue
    libva-2.21.0-2.fc40 doesn't appear in repo despite been tagged into koji f40.

It seems because the older bohdi update was merged into stable after the newer one:
https://bodhi.fedoraproject.org/updates/FEDORA-2024-5cc15931d7 masking the newer libva-2.21.0-2.fc40

koji untag-build f40 libva-2.21.0-1.fc40
or
Re-applying the https://bodhi.fedoraproject.org/updates/FEDORA-2024-5cc15931d7 bodhi update will fix the issue.

  • When do you need this? (YYYY/MM/DD)
    next compose

  • When is this no longer needed or useful? (YYYY/MM/DD)
    still useful

  • If we cannot complete your request, what is the impact?
    older package will not be published.


Metadata Update from @jnsamyak:
- Issue assigned to jnsamyak

Metadata Update from @jnsamyak:
- Issue tagged with: low-gain, low-trouble, ops

I've submitted another bodhi update to avoid missing the next compose (and more importantly the f40 final freeze).

https://bodhi.fedoraproject.org/updates/FEDORA-2024-637efc913f

I ran our check script and it came up with:

        calendar-1.37-6.20240209cvs.fc40 < calendar-1.37-7.20211220cvs.fc37
        cura-fdm-materials-5.4.0-2.fc40 < cura-fdm-materials-5.6.0-3.fc40
        golang-github-nats-io-jwt-1.2.2-4.fc40 < golang-github-nats-io-jwt-1.2.2-6.fc38
        golang-google-grpc-1.48.0-1.fc40 < golang-google-grpc-1.48.0-9.fc39
        libva-2.21.0-1.fc40 < libva-2.21.0-2.fc40
        patool-1.12-0.27.20231014gitab64562.fc40 < patool-1.12-22.fc39
        pgadmin4-8.4-1.fc40 < pgadmin4-8.4-2.fc40
        python-pywlroots-0.16.4-3.fc40 < python-pywlroots-0.16.6-1.fc40
        resalloc-5.3-1.fc40 < resalloc-5.4-1.fc40
        vim-devicons-0.11.0-0.20221001git71f239a.13.fc40 < vim-devicons-0.11.0-10.20200509gitd12c9b4.fc39

Of those, calendar, patool, and vim-devicons seem like packaging mistakes/upgrade path problems and we should not fix those, but get maintainers to fix them

The rest the newer one can be retagged.

calendar: https://bugzilla.redhat.com/show_bug.cgi?id=2271965
vim-devicons: https://bugzilla.redhat.com/show_bug.cgi?id=2271966
patool looks like rpmdev-bumpspec used in the f39 mass rebuild did the wrong thing and bumped it to a non desired version... it's fine it's not latest.

I'll do another run tomorrow and retag things (I tried now, but updates pushes are happening and it messed up things)

I did ran the script this morning:

)   calendar-1.37-6.20240209cvs.fc40 < calendar-1.37-7.20211220cvs.fc37
    cura-fdm-materials-5.4.0-2.fc40 < cura-fdm-materials-5.6.0-3.fc40
    golang-github-nats-io-jwt-1.2.2-4.fc40 < golang-github-nats-io-jwt-1.2.2-6.fc38
    golang-google-grpc-1.48.0-1.fc40 < golang-google-grpc-1.48.0-9.fc39
    libva-2.21.0-1.fc40 < libva-2.21.0-2.fc40
    patool-1.12-0.27.20231014gitab64562.fc40 < patool-1.12-22.fc39
    pgadmin4-8.4-1.fc40 < pgadmin4-8.4-2.fc40
    python-pywlroots-0.16.4-3.fc40 < python-pywlroots-0.16.6-1.fc40
    resalloc-5.3-1.fc40 < resalloc-5.4-1.fc40
    vim-devicons-0.11.0-0.20221001git71f239a.13.fc40 < vim-devicons-0.11.0-10.20200509gitd12c9b4.fc39

where i found out these, so following needs to be retagged if we ignore the above mentioned packages

cura-fdm-materials
golang-github-nats-io-jwt
golang-google-grpc
pgadmin4
python-pywlroots
resalloc

golang-github-nats-to-jwt is another broken update path. ;(
as is golang-google-grpc

Can you file bugs on those 2? or would you like me to?

So that leaves us with:

First make sure they are all signed:

koji write-signed-rpm e99d6ad1 cura-fdm-materials-5.6.0-3.fc40 libva-2.21.0-2.fc40 pgadmin4-8.4-2.fc40 python-pywlroots-0.16.6-1.fc40 resalloc-5.4-1.fc40

Then retag them in:
koji tag-pkg f40 --force cura-fdm-materials-5.6.0-3.fc40 libva-2.21.0-2.fc40 pgadmin4-8.4-2.fc40 python-pywlroots-0.16.6-1.fc40 resalloc-5.4-1.fc40

libva-2.21.0-3.fc40 is tagged in f40 i dont think we need to that now?

I retagged them all but cura-fdm-materials-5.6.0-3.fc40 threw a generic error while running write-signed-rpm on it

koji write-signed-rpm e99d6ad1 cura-fdm-materials-5.6.0-3.fc40
[1/2] cura-fdm-materials-5.6.0-3.fc40.src
[2/2] cura-fdm-materials-5.6.0-3.fc40.noarch
2024-03-29 10:04:57,440 [ERROR] koji: GenericError: No cached signature for package cura-fdm-materials-5.6.0-3.fc40.src, key e99d6ad1

And, I left libva as well since a newer bodhi update went into f40

I'll do the bug filing now for the rest two!

Why it went backward ? I have purposely issued another bodhi update for libva- 2.21.0-3 to avoid issue with the -1 version still in repo.
https://mirror.karneval.cz/pub/linux/fedora/linux/development/40/Everything/x86_64/os/Packages/l/libva-2.21.0-1.fc40.x86_64.rpm

Can you please untag the -1 build, I have a dependency failing to build for 2 weeks because of that.
koji untag-build f40 libva-2.21.0-1.fc40

At least I don't see anything wrong on the koji side, so bodhi is probably messing here.
$ koji list-history --tag f40 --package libva
Tue Aug 8 18:46:01 2023 package owner kwizart set for libva in f40 by humaton [still active]
Tue Aug 8 18:46:01 2023 package list entry created: libva in f40 by humaton [still active]
Tue Aug 8 18:55:03 2023 libva-2.19.0-2.fc39 tagged into f40 by humaton
Sun Sep 17 17:02:45 2023 libva-2.20.0-1.fc40 tagged into f40 by bodhi
Tue Jan 16 23:51:26 2024 libva-2.20.0-2.fc40 tagged into f40 by bodhi [still active]
Mon Jan 29 17:37:35 2024 libva-2.20.0-4.fc40 tagged into f40 by releng [still active]
Wed Jan 31 15:58:44 2024 libva-2.19.0-2.fc39 untagged from f40 by oscar
Thu Mar 7 23:23:17 2024 libva-2.20.0-6.fc40 tagged into f40 by bodhi [still active]
Sat Mar 9 17:13:36 2024 libva-2.20.0-1.fc40 untagged from f40 by oscar
Sat Mar 23 01:16:21 2024 libva-2.21.0-2.fc40 tagged into f40 by bodhi [still active]
Sat Mar 23 01:16:21 2024 libva-2.21.0-1.fc40 tagged into f40 by bodhi [still active]
Fri Mar 29 01:15:30 2024 libva-2.21.0-3.fc40 tagged into f40 by bodhi [still active]

Okay, it might be because the repo generation failed for today (29/03). Only the repo created from 28/03 is live on current repositories...

Hopefully this can be fixed before the Final freeze...

Yes, the compose failed. Another is running now. ;)

Should be all sorted out now?

Please re-open if you still see any issues...

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

Metadata