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
Consolatis has quit [Ping timeout: 480 seconds]
Consolatis has joined #freedesktop
scrumplex_ has joined #freedesktop
scrumplex has quit [Ping timeout: 480 seconds]
scrumplex has joined #freedesktop
scrumplex_ has quit [Ping timeout: 480 seconds]
ity1 has joined #freedesktop
ity has quit [Remote host closed the connection]
Consolatis_ has joined #freedesktop
Consolatis is now known as Guest11840
Consolatis_ is now known as Consolatis
Guest11840 has quit [Ping timeout: 480 seconds]
Xe_ has joined #freedesktop
alanc has quit [Remote host closed the connection]
alanc has joined #freedesktop
Xe has quit [Ping timeout: 480 seconds]
MrCooper_ has joined #freedesktop
MrCooper has quit [Ping timeout: 480 seconds]
noodlez1232 has quit [Remote host closed the connection]
noodlez1232 has joined #freedesktop
krushia has quit [Ping timeout: 480 seconds]
MrCooper__ has joined #freedesktop
MrCooper_ has quit [Ping timeout: 480 seconds]
gnuiyl has quit [Remote host closed the connection]
gnuiyl has joined #freedesktop
swatish2 has joined #freedesktop
konstantin_ has joined #freedesktop
konstantin is now known as Guest11845
konstantin_ is now known as konstantin
Guest11845 has quit [Read error: Connection reset by peer]
MrCooper_ has joined #freedesktop
AbleBacon has quit [Read error: Connection reset by peer]
MrCooper__ has quit [Ping timeout: 480 seconds]
ximion has quit [Remote host closed the connection]
jsa1 has joined #freedesktop
sghuge has quit [Remote host closed the connection]
sghuge has joined #freedesktop
sima has joined #freedesktop
tzimmermann has joined #freedesktop
swatish21 has joined #freedesktop
swatish2 has quit [Ping timeout: 480 seconds]
swatish2 has joined #freedesktop
MrCooper_ is now known as MrCooper
swatish21 has quit [Ping timeout: 480 seconds]
dcunit3d has quit [Quit: No Ping reply in 180 seconds.]
dcunit3d has joined #freedesktop
swatish2 has quit [Ping timeout: 480 seconds]
swatish2 has joined #freedesktop
krei-se- has joined #freedesktop
krei-se has quit [Ping timeout: 480 seconds]
<bentiss> daniels: so far, I got the fleeting runners up, but I wonder, do we want them to be privileged? Do we want kvm support for them (I don't think we can actually)? How can we test them? (little help here) and Should we request for limit bump from Hetzner because we are already being capped by the server count and dedicated CPU count
<bentiss> For anyone will to test them (not privileged and no kvm ATM), use the tag 'hetzner-test-x86' and 'hetzner-test-aarch64'
<bentiss> For anyone *willing
JanC is now known as Guest11857
JanC has joined #freedesktop
Guest11857 has quit [Read error: Connection reset by peer]
mvlad has joined #freedesktop
jsa1 has quit [Ping timeout: 480 seconds]
swatish2 has quit [Ping timeout: 480 seconds]
fantom has joined #freedesktop
jsa1 has joined #freedesktop
swatish2 has joined #freedesktop
mattst88_ has joined #freedesktop
mattst88 has quit [Read error: Connection reset by peer]
Abbott has joined #freedesktop
<Abbott> the window switcher on openbox and on rofi is not picking up the discord icon. I can see there is a 256x256 icon installed at /usr/share/icons/hicolor/256x256/apps/discord.png but that is not being used, I guess. Is there another size that xdg generally expects to have that the window switcher could be expecting and not finding?
guludo has joined #freedesktop
swatish2 has quit [Ping timeout: 480 seconds]
<pq> bentiss, I fired a test pipeline, no clue what the token failure is: https://gitlab.freedesktop.org/pq/fourbyfour/-/jobs/73107169
<pq> the ci-fairy job in the same pipeline ran fine
<bentiss> pq: thanks for testing!
* bentiss looks into that failure
<bentiss> should be easy enough to reproduce with just `podman login -u $CI_REGISTRY_USER -p $CI_REGISTRY_PASSWORD $CI_REGISTRY`
<pq> all that comes from ci-templates
<bentiss> yeah
<bentiss> I wonder if this is because the runner is not privileged
<bentiss> pq: so definitely making the runners privileged is much better for podman login
<bentiss> so I guess that solves my first question :)
<bentiss> pq: I've restarted your job to see if this goes further
<pq> cool
<pq> haha, 'reuse' hilarity, but otherwise CI seems to work :-D
<pq> that is my problem
rgallaispou has joined #freedesktop
<bentiss> yeah, I was starting to wonder if this was an issue on my side :)
<pq> MR page "checking pipeline status" is spinning a long time.
<pq> page reload made it appear fast
<pq> the whole pipeline passed now
<bentiss> yeah, the background jobs have a little bit of lag because we are pruning all the mesa pipelines older than a year
<pq> good good :-)
<bentiss> and there is a waiting queue of about 2 minutes from time to time
<pq> Once the dust settles, some summary guide on recommended clean-up settings and rules would be nice. Or is it enough to use ci-templates' FDO_EXPIRES_AFTER? I guess artifacts and(?) pipelines would need some clean-up settings?
<bentiss> artifacts will be purged after a year, no matter what (policy at the bucket level)
<bentiss> for pipelines, there is a settings you can set, but in the same way I'll probably just enforce it for any project (settings->ci/cd->general->Automatic pipeline cleanup)
krushia has joined #freedesktop
<bentiss> FDO_EXPIRES_AFTER for the container image is nice also but can't work for upstream projects pipelines (because you want to keep those around)
GNUmoon2 has quit [Remote host closed the connection]
GNUmoon2 has joined #freedesktop
<bentiss> daniels: we got our server limit upgraded from 20 to 50. Did you do anything?
<mupuf> bentiss: what is that supposed to mean?
<bentiss> mupuf: what's your question? what is what supposed to mean?
<mupuf> bentiss: "we got our server limit upgraded from 20 to 50"
<bentiss> oh
<bentiss> well, hetzner cloud limits the number of servers you can request on the cloud (probably to prevent you to have a too big bill you can not pay)
<bentiss> and we were having a limit of 20, and now we are at 50
<bentiss> which is good, because then I can fire more servers for the runners (one per job, so I don't have to mess with the cpu map), instead of having bigger servers with multiple jobs on each
<mupuf> Ahhh, I see
<mupuf> nice, thanks
rgallaispou has quit [Read error: Connection reset by peer]
rgallaispou has joined #freedesktop
alatiera2 has joined #freedesktop
alatiera has quit [Read error: Connection reset by peer]
alatiera2 is now known as alatiera
<daniels> bentiss: I put the request in when we first signed up, but nothing beyond that
<bentiss> k, well, maybe it's a "1 month anniversary" feature :)
<bentiss> anyway. I've put in prod a beaffy dual shared-cpu for placeholder jobs, each can have 100 jobs, with an autoscaling feature of 2.
<bentiss> I know, this is impressive (and I ǘe disabled the ones on Equinix)
<daniels> bentiss: I wonder why podman login requires privileged mode though ... ? I thought that wasn't the case at all, and kind of sucks if it is
<bentiss> it's probably the way we configured it in the container
<bentiss> IIRC gnome manages to run without privileged runners
<daniels> yeah, I can try to find out from them if they had to do anything special
<bentiss> but OTOH, now except for the kvm ones, we have one VM per job (which can be reused), so does it really matters?
<daniels> the KVM thing kind of sucks though as we definitely need those for testing Weston, libinput, DRM/VKMS, Mesa, etc etc
<bentiss> well, I was thinking that instead of having 3 equivalent machines we had on Equinix, we could have just one, and make it run only kvm
<bentiss> rest will go through cloud with the autoscaler feature
<bentiss> arguably, cloud is twice as expensive, but with autoscaler we can hope for something lower
<bentiss> the only thing I don't know how to set up in the cloud VM is the ccache we had, because the VMs are ephemeral
<daniels> oh yeah, that's a really good point actually - I like it!
<bentiss> and the cache feature of gitlab (https://docs.gitlab.com/ci/caching/) is not exactly that
<daniels> meh, last time we tried any kind of shared cache, it turned out to be way slower to copy the data than to recompile :P
<bentiss> heh
<bentiss> I guess I can try redeploying the x86 and arm and let them handle normal jobs (not just the tagged ones)
<bentiss> though it's friday afternoon...
<karolherbst> can somebody remove the abuse report 3439? It's stale
<karolherbst> or rather.. in a state nobody can do something about it from the UI
haaninjo has joined #freedesktop
<pq> bentiss, oh, can I remove the hetzner-test-x86 tag already?
rgallaispou has quit [Read error: Connection reset by peer]
<bentiss> pq: yeah, we'll see at scale how that's behaving :)
<bentiss> pq: thanks a lot for testing!
<pq> Thank you!
rgallaispou has joined #freedesktop
<bentiss> oh... grafana.freedesktop.org is not deployed (looking at failed job https://gitlab.freedesktop.org/gfx-ci/mesa-performance-tracking/-/jobs/73113607)
<eric_engestrom> bentiss: in the mesa ci, I'm getting 413 Request Entity Too Large for the mesa build (150MB), is that expected?
<eric_engestrom> + curl --header 'Authorization: Bearer [MASKED]' -X PUT --form file=@mesa-arm32-default-debugoptimized.tar.zst https://s3.freedesktop.org/artifacts/eric/mesa/1386181/
<bentiss> eric_engestrom: yeah, I saw that... don't have an answer for it ATM
<bentiss> (I must confess that my brain disconnected since Wednesday when I realized the db was fine in the end)
<eric_engestrom> ack; is that something you'll be able to look into? I'm really struggling for time and I want devs to be able to start using the CI again ASAP
<eric_engestrom> re- "db is fine in the end", does that mean we are no longer in "might have to revert, consider it read-only" mode?
<bentiss> one thing we could do is use multipart archives
<bentiss> eric_engestrom: yeah
<eric_engestrom> "multipart archives" -> that would require a lot of infra changes, I'd rather avoid that
<bentiss> k
ximion has joined #freedesktop
mripard has quit [Quit: WeeChat 4.5.2]
<bentiss> good news: must be coming from our settings: https://github.com/oxyno-zeta/s3-proxy/issues/383 "The fact is that I upload files larger than 100GB, sometimes 300GB each." -> this seems possible
<bentiss> could be nginx related
jsa1 has quit [Ping timeout: 480 seconds]
* bentiss can't remember which settings we had in nginx to allow for bigger uploads :(
<bentiss> seems like it was in the ingress settings
<eric_engestrom> bentiss: I started a new pipeline, should help with the missing images and it has some fixes I had locally
<bentiss> eric_engestrom: I'll be afk a bit (school pick up)
<eric_engestrom> (I configured my registry to delete everything after one day, which is why your retry of a single job failed)
<bentiss> eric_engestrom: but I've copied the ingress settings we have in gitlab, that might help
<bentiss> kk
<eric_engestrom> thanks! I'll let you know if a bit if that worked :)
rgallaispou has left #freedesktop [#freedesktop]
webreader has joined #freedesktop
webreader has quit [Quit: Page closed]
<eric_engestrom> bentiss: your fix worked!
<eric_engestrom> I'll need to do a full check, but I think we're ready to re-enable mesa ci :)
<bentiss> \o/
* bentiss just needs to update the deployment definition then
<bentiss> and done
<eric_engestrom> thanks for your help!
<eric_engestrom> is the plan to fix s3cp as well, or to drop it and have all projects move to using curl directly?
<bentiss> curl seems like a good option
<bentiss> no?
<bentiss> if it works
<eric_engestrom> yeah, imo it's simpler
<bentiss> plus s3-proxy has support for plain oidc, so you can even use your browser to download the file you're allowed to, while before you had to use token + PITA
<eric_engestrom> the downside is that you have to copy/paste more everywhere, which is a hassle when you have to update it, but the upside is that there's no delay to fix things, no need to wait for one repo to have the fix so that other repos can update
<bentiss> so less technical debt is better ;)
<bentiss> yeah, plus the potential impact of bumping the ci-fairy include sha which may or may not break other ci-fairy commands depending on the branch
<eric_engestrom> yep
<eric_engestrom> pipelines are not being created though, guessing the background tasks are busy deleting old pipelines instead? ^^
<bentiss> yeah... 4 min of queue
<bentiss> I've added one sidekiq worker, though I hope I won't have teh PG memory issues
<bentiss> and dropped it, now that the heat is gone
<bentiss> week end for me. If anything wrong happens with the runner I hope one gitlab admin will just pause the 2 new in the gitlab runner interface
swatish2 has joined #freedesktop
* mupuf can be pinged if the above happens ^
<mupuf> thq
<mupuf> thanks a lot for everything, bentiss:!
<mupuf> Enjoy your well-deserved weekend
tzimmermann has quit [Quit: Leaving]
webreader has joined #freedesktop
<kxkamil> there is still 'consider this as read-only' - is it obsolate?
<kxkamil> is it working: issue create/change label/close, addind a user as reporter/dev?
swatish2 has quit [Ping timeout: 480 seconds]
<DragoonAethis> kxkamil: Everything works except GitLab CI (runner setup in progress), some artifacts (releases, uploads, etc) and container registries
<DragoonAethis> Git itself, issues, user management, etc works
AbleBacon has joined #freedesktop
<kxkamil> great, thx all, esp bentiss!
nephyrin_ has quit [Quit: ... besides, it was hot]
nephyrin has joined #freedesktop
nephyrin has quit [Quit: ... besides, it was hot]
tanty has quit [Quit: Ciao!]
mvlad has quit [Remote host closed the connection]
nephyrin has joined #freedesktop
haaninjo has quit [Quit: Ex-Chat]
imre has quit [Ping timeout: 480 seconds]
<eric_engestrom> placeholder-job jobs are all hitting "no space left on device"
tanty has joined #freedesktop