ChanServ changed the topic of #freedesktop to: https://www.freedesktop.org infrastructure and online services || for questions about freedesktop.org projects, please see each project's contact || for discussions about specifications, please use https://gitlab.freedesktop.org/xdg or xdg@lists.freedesktop.org
mohamexiety has quit []
agd5f_ has joined #freedesktop
agd5f has quit [Ping timeout: 480 seconds]
agd5f has joined #freedesktop
agd5f_ has quit [Ping timeout: 480 seconds]
co1umbarius has joined #freedesktop
columbarius has quit [Ping timeout: 480 seconds]
ybogdano has joined #freedesktop
ximion has quit [Quit: Detached from the Matrix]
BLumia[m] has joined #freedesktop
<BLumia[m]>
Hi, I have a MR on xdg-spec repo which I don't know how to get it merged, is there the correct place to request a review or discuss it?
jarthur has quit [Read error: Connection reset by peer]
thaller has joined #freedesktop
alanc has quit [Remote host closed the connection]
alanc has joined #freedesktop
jarthur has joined #freedesktop
damian_ has quit [Ping timeout: 480 seconds]
damian has joined #freedesktop
ofourdan has joined #freedesktop
AbleBacon has quit [Read error: Connection reset by peer]
thaller has quit [Remote host closed the connection]
thaller has joined #freedesktop
kbingham has joined #freedesktop
feto_bastardo has quit [Quit: quit]
feto_bastardo has joined #freedesktop
mohamexiety has joined #freedesktop
<eric_engestrom>
PSA: igalia farm is running at half capacity, one of the runners is down
<eric_engestrom>
BLumia[m]: see /topic, this channel is about the infrastructure on freedesktop.org, not the projects hosted there
<eric_engestrom>
BLumia[m]: I would suggest sending an email to xdg@lists.freedesktop.org if you want to get more people from that project aware of your MR :)
<BLumia[m]>
Oh sorry the topic is collapsed by default in my client. Thanks for the info :)
robobub has quit []
pendingchaos_ is now known as pendingchaos
<eric_engestrom>
no worries, we get this quite often ^^
vkareh has joined #freedesktop
<eric_engestrom>
igalia farm back at 100% :)
<eric_engestrom>
(and immediately that one user takes all the runners again)
<eric_engestrom>
we really need to have documentation on ci_run_n_monitor.py to point people to so that they stop DoS'ing everything when they want to run one test
<eric_engestrom>
(this particular one is making zink changes but running the entire CI every time)
<mupuf>
eric_engestrom: lol
<mupuf>
yeah... maybe we should force people to say which driver they want to work on...
<mupuf>
but then it really feels like re-implementing part of gitlab's "changes"
mohamed has joined #freedesktop
ximion has joined #freedesktop
mohamexiety has quit [Ping timeout: 480 seconds]
mohamed has quit []
agd5f_ has joined #freedesktop
agd5f has quit [Ping timeout: 480 seconds]
agd5f has joined #freedesktop
agd5f_ has quit [Ping timeout: 480 seconds]
iNKa is now known as Brocker
agd5f_ has joined #freedesktop
agd5f has quit [Ping timeout: 480 seconds]
agd5f has joined #freedesktop
<mupuf>
bentiss: is it meant for fdo-equinix-m3l-19 to be offline?
<bentiss>
mupuf: nope
agd5f_ has quit [Ping timeout: 480 seconds]
<mupuf>
bentiss: that may be the reason why the sanity job took over 13 minutes to even run :D
<bentiss>
mupuf: yep. I can not ssh on to it, and the console is not responding
<bentiss>
we had that last week, and I had to resort to kill the server
<bentiss>
I wonder what happened, there were no suspicious jobs AFAICT
<mupuf>
lovely
<bentiss>
yeah, it is dead: Power operation failed
Leopold__ has quit [Remote host closed the connection]
Leopold_ has joined #freedesktop
<DavidHeidelberg[m]>
eric_engestrom: at least for lava, we have priorities, can't you do something like that too?
<DavidHeidelberg[m]>
so the general audience is prioritized lower than Marge
<eric_engestrom>
DavidHeidelberg[m]: that sounds desirable, but I have no idea how this could be done
<mupuf>
eric_engestrom: you are using the baremetal infra, right?
<eric_engestrom>
as far as I know we just get whatever gitlab gives us
<eric_engestrom>
mupuf: yeah
<mupuf>
yeah, then until I am done with hacking gitlab, you can't do anything...
<eric_engestrom>
gitlab has an issue somewhere tracking the addition of the concept of job priority, but it's not there yet
<DavidHeidelberg[m]>
eric_engestrom: on lava, there is queue and when user job wating and also Marge job waiting, even if Marge is sumbitted later, it's procesed first
<mupuf>
DavidHeidelberg[m]: yeah, but it is still shit though, as jobs are likely to time out because gitlab doesn't know it is waiting on a machine
<DavidHeidelberg[m]>
but you would have to put something between bm and gitlab
<DavidHeidelberg[m]>
yeah, there has to be middle man.. ehm, computer...
<bentiss>
mupuf: m3l-19 is now gone, and m3l-20 is handling workloads
<eric_engestrom>
"this currently is not prioritized"
<mupuf>
but yeah, when I come back, I guess I'll deploy gitlab omnibus and start hacking on the job prioritization support
<mupuf>
eric_engestrom: that's not what you want
<mupuf>
that's nice, but that's not what you want
<mupuf>
this is just to prioritize jobs within an pipeline
<mupuf>
(run this job first, not this one)
<mupuf>
what we need is for runners to be able to query the list of jobs it could run, then pick the one it wants to run first
<eric_engestrom>
wouldn't it be across all the jobs currently waiting, regardless of pipeline?
<mupuf>
nope, and that would be silly to try as it would be collaborative between all the projects :D
<eric_engestrom>
yeah if the runner can pick that's more flexible, but also more complex to implement I assume
<mupuf>
I doubt it would
<eric_engestrom>
ok
<mupuf>
right now, runners just ask "Give me a job"
<mupuf>
and no possibility of refusing it
<eric_engestrom>
and yeah, collaborating with all the projects on the instance can't work, but I was thinking within a namespace (including its fork network)
<mupuf>
so... just add 2 endpoints: Give me the list of jobs, and "I'll pick this one if it is still available"
<mupuf>
gotta go!
<eric_engestrom>
o/
<eric_engestrom>
yeah that sounds like a better design, but obviously it opens to race conditions; not sure how much of a problem that could be for someone somewhere though
agd5f_ has joined #freedesktop
agd5f has quit [Ping timeout: 480 seconds]
agd5f has joined #freedesktop
agd5f_ has quit [Ping timeout: 480 seconds]
agd5f_ has joined #freedesktop
Haaninjo has joined #freedesktop
agd5f has quit [Ping timeout: 480 seconds]
agd5f has joined #freedesktop
agd5f_ has quit [Ping timeout: 480 seconds]
<DavidHeidelberg[m]>
bentiss: can I get marge-bot group on FDO for our marge-bot fork?
<bentiss>
DavidHeidelberg[m]: -EPARSE
<DavidHeidelberg[m]>
hmm, thinking about the name. user is called marge-bot
<alatiera>
DavidHeidelberg[m] the gnome fork is one of my branches with like 2 patches on top 😆
<alatiera>
and it wasn't until recently that I did it
<DavidHeidelberg[m]>
alatiera: what you think about hosting it on FDO? Does it snd like good idea/
<DavidHeidelberg[m]>
?
<DavidHeidelberg[m]>
I perceive as FDO as a shared playground for most of desktop FOSS projects, spec, base stones...
<alatiera>
DavidHeidelberg[m] sure thing
<__tim>
DavidHeidelberg[m], why "new upstream"? I mean, is it an fdo-specific fork or replacing an existing defunct upstream (with their consent, somewhat)
<alatiera>
the second one
<DavidHeidelberg[m]>
__tim: replacing upstream repo (we would do it as a specific fork.. maybe some specific branch for Mesa3D), but upstream is dead
<DavidHeidelberg[m]>
in mesa3d repo we hit some mental breakdowns of Marge-bot, so we need to fix it + possibily improve it, but preferably in some most upstream way (if we won't be alone doing all heavy lifting)
<bentiss>
alatiera: done. Should be deployed. Hopefully. maybe :)
<alatiera>
bentiss thanks!
<bentiss>
no worries
robobub has joined #freedesktop
<alatiera>
ping me if there's still egress from ci images
<__tim>
I think it was fairly easy to spot with the cerbero trigger jobs which would always/often re-download the image