26 Jul 2024 |
Zlopez | You should be able to get python backtrace by using py-bt | 17:10:06 |
jednorozec | I will be around to fix things during the weekend | 17:36:29 |
Zlopez | We can just block the offending toddlers on staging and fix the things on Monday | 17:40:57 |
Zlopez | The changes are now deployed on staging | 17:50:25 |
Zlopez | No error during deployment | 17:51:50 |
Zlopez | And I'm logging out for today | 17:51:58 |
Zlopez | Bye everyone | 17:52:07 |
nirik | night Zlopez. Have a great weekend. | 18:31:38 |
jednorozec | ha now I finally have the correct data for the bz sync toddler | 18:53:46 |
jednorozec | nirik: so what about I deploy the new toddlers and disable pdc in prod on monday morning local time. By the time you are online everything should be just enough under fire | 18:55:35 |
nirik | alright. works for me. | 18:57:15 |
nirik | Jeremy Cline: added some stack traces to https://pagure.io/sigul/issue/25 if you can make anything of em... | 22:01:39 |
Jeremy Cline | Mm, if you are still attached does py-bt work? | 22:03:32 |
Jeremy Cline | In reply to @nirik:matrix.scrye.com Jeremy Cline: added some stack traces to https://pagure.io/sigul/issue/25 if you can make anything of em... https://wiki.python.org/moin/DebuggingWithGdb has instructions (valid as of centos 7) on setting up the python extension for gdb if py-bt doesn't work | 22:09:19 |
nirik | I'm not, but it does not. | 22:11:20 |
Carl George | nirik: think i could pick your brain for a few minutes about bodhi stuff? | 22:14:09 |
nirik | sure, but trying to debug sigul, so could it wait a bit? | 22:14:27 |
Carl George | of course | 22:14:39 |
Carl George | i'll just type it up here and you can reply whenever you're ready so we don't play too much tag | 22:14:59 |
Carl George | i realize that we never updated get_active_releases_from_bodhi() in owner-sync-pagure to account for epelX-next branches.
https://pagure.io/fedora-infra/ansible/blob/main/f/roles/bodhi2/backend/templates/owner-sync-pagure.j2
i'm not aware of any negative effects from missing that. if that's fine, do you think it will be fine only care about syncing ownership with the epel10 branch, and having this script ignore branches like epel10.0 and epel10.1? in theory there shouldn't be any difference in ownership between epel10 and epel10.0, just like in theory there shouldn't be any between epel9 and epel9-next. | 22:19:50 |
Jeremy Cline | In reply to @nirik:matrix.scrye.com I'm not, but it does not. Ah, it starts working automagically if you install python3-debuginfo I think | 22:27:44 |
nirik | yes, figured that. just spent a while getting the stupid debuginfo installed. ;) | 22:32:08 |
nirik | it's installed and gdb is attached, so next time I can just control -c and py-bt it. | 22:32:26 |
nirik | Carl George: if ownership isn't going to be different per branch, and if epel10.x inherits from epel10, yeah, no need to sync that. It's largely unused anyhow... I think the fails to build script uses it? or something small like that... | 22:33:57 |
Carl George | yeah i'm not 100% clear on the value of this script. there is just one owner in koji, right? and there is only one main admin in pagure, so we shouldn't ever have variance between branches, even between fedora and epel. the only difference would be if we were accounting for collaborator perms for just epel, but that would require koji having a concept of multiple owners. | 22:37:16 |
Carl George | let me know if i'm off base on any of that | 22:37:23 |
nirik | yep.. it's really for some corner cases and we should probibly find those and fix them anyhow. | 23:32:31 |
Carl George | i'm gonna spend some more time wrapping my head around this script so i can make sure we have the necessary epel10 adjustments | 23:38:24 |
Carl George | breaking the assumption of tag name == branch name is going to have some interesting side effects, in this script and in other places | 23:39:09 |
nirik | yeah | 23:44:41 |