02:16
ngcortes has quit [Remote host closed the connection]
02:40
wirez has joined #freedesktop
04:56
bengal has joined #freedesktop
05:09
danvet has joined #freedesktop
05:57
wirez has quit [Ping timeout: 480 seconds]
06:14
Yeldham has joined #freedesktop
06:21
bcarvalho has quit [Remote host closed the connection]
06:31
alanc has quit [Remote host closed the connection]
06:31
alanc has joined #freedesktop
07:30
bengal has quit [Remote host closed the connection]
07:30
bengal has joined #freedesktop
07:39
xexaxo has joined #freedesktop
07:40
sumits has joined #freedesktop
07:42
bengal has joined #freedesktop
07:50
bcarvalho has joined #freedesktop
07:53
xexaxo has quit [Ping timeout: 480 seconds]
08:14
sumits has joined #freedesktop
08:17
bengal has joined #freedesktop
09:00
K`den has joined #freedesktop
09:00
Kayden has quit [Read error: Connection reset by peer]
09:05
bengal has joined #freedesktop
09:10
xexaxo has joined #freedesktop
09:53
xexaxo has quit [Remote host closed the connection]
09:53
xexaxo has joined #freedesktop
11:28
reillybrogan has joined #freedesktop
11:33
reillybrogan_ has quit [Ping timeout: 480 seconds]
11:40
Adrinael has joined #freedesktop
11:43
<
Adrinael >
Getting test timeouts on gitlab runners, at least fdo-packet-m1xl-1 fdo-packet-m1xl-3
11:44
<
Adrinael >
Also some earlier on fdo-packet-m1xl-2
11:44
<
Adrinael >
daniels, ^
11:46
<
daniels >
Adrinael: please provide job links
12:30
<
tomeu >
those are quite weird
12:30
<
tomeu >
shadeslayer: did you retry any jobs in that pipeline?
12:31
<
tomeu >
maybe those are related with the timeouts reported just above
12:32
<
tomeu >
eg. the runner took so long to pick up a job, that the token had already expired by then
12:32
<
tomeu >
maybe we should make the TTL of the tokens a bit higher
12:44
<
daniels >
that was transient whilst I was dealing with some breakage
12:44
<
daniels >
retry and it'll be fine
12:44
<
daniels >
I was hoping no-one would catch the 10-15min window where nothing worked :P
12:45
<
shadeslayer >
oops :P
12:46
<
bentiss >
daniels: out of curiosity, what was broken?
12:59
<
bentiss >
sigh large-5 is now down, ceph is timeing out
13:03
<
daniels >
bentiss: k3s-minio-1 was logging a ton of errors about how the master's cert was expired so the node wasn't polling for updates
13:03
<
bentiss >
daniels: oh, ok
13:04
<
daniels >
so the opa pod was still running, but when I deleted it (or scaled the rs/deploy) it wasn't being recreated
13:04
<
bentiss >
daniels: IIRC you can solve that error by restarting k3s
13:04
<
daniels >
bentiss: exactly!
13:12
<
bentiss >
daniels: m1xl-3 is having quite some load
13:12
<
bentiss >
Load average: 3222.30
13:12
<
bentiss >
lots of virgl_test_server --use-egl-surfaceless
13:12
<
daniels >
sweet jesus
13:12
<
daniels >
how many of them?
13:13
<
tomeu >
ouch, that already happened once
13:13
<
tomeu >
do we know if it's from mesa or virglrenderer's CI?
13:13
<
bentiss >
also we have quite some docker pods still running for the past 43h
13:15
<
bentiss >
daniels: I wonder if that is not jobs that have not been killed in the background
13:17
<
bentiss >
I count 1818 [virgl_test_serv] on m1xl-3
13:20
<
bentiss >
I paused m1xl-3, once its jobs are pruned, I'll killall virgl_test_serv
13:20
<
daniels >
ouch yeah, m1xl-3 is utterly slammed with 3492 load avg and all CPUs 100% utilised and a lot in D state, otoh m1xl-2 has 1400 load avg but CPUs (whilst virgl-test was running anyway, it's finished now) mostly idle ... ?
13:21
<
bentiss >
(side note, why can't I lock m1xl-3?)
13:22
ezequielg has quit []
13:22
ezequielg has joined #freedesktop
13:23
<
bentiss >
Unable to get a valid fd
13:26
<
daniels >
yeah I’d just kill it
13:26
<
daniels >
maybe that’s the issue then - we’re stuck in infinite coredump hell?
13:27
<
bentiss >
coredump hell or fork bomb, don't know
13:27
<
bentiss >
using docker kill -s 9 CONTAINER_ID is not working
13:27
wyre_ has left #freedesktop [#freedesktop]
13:28
wyre has joined #freedesktop
13:28
<
bentiss >
I guess a reboot of m1xl-3 is in order
13:29
<
bentiss >
hard reboot seems like my last option :/
14:19
<
daniels >
bentiss: I think my first attempt at disabling core dumps was broken, so I've tried again on all of the shared runners
14:19
<
daniels >
I don't think it's forkbomb
14:19
<
daniels >
and every single dEQP coredumping would definitely account for coredump hell
14:19
<
daniels >
so fingers crossed that does the trick ... ?
14:19
<
daniels >
Adrinael: ^ that should solve your problem I hope
14:24
<
daniels >
there are 1675 igt tasks on packet-m1xl-1 in D state ...
14:25
<
daniels >
can you please make sure you're using meson test with the --parallel arg as ${FDO_CI_CONCURRENT:-8} so it doesn't try to execute 48 igts in parallel?
14:26
<
daniels >
(and similarly with -j for your build jobs)
14:34
<
daniels >
yeah ok, igt is DoSing itself
14:35
<
Adrinael >
gah patch coming up
14:36
<
daniels >
I've already sent one to the list
14:36
<
daniels >
oh no I haven't
14:38
MrCooper has quit [Quit: Leaving]
14:38
<
daniels >
there we are, fixed git-send-email
14:40
MrCooper has joined #freedesktop
14:41
<
daniels >
(think it's stuck in igt-dev@ moderation ... having actual MRs for igt sure would be nice tbh)
14:42
<
Adrinael >
Hopefully coming in the future, but not in the near future
14:42
<
MrCooper >
sure would be
14:48
<
daniels >
oops, patch is broken, v2 incoming
14:49
<
Adrinael >
My first one was also broken =(
15:19
<
Adrinael >
hmm that's the only test job in this pipeline that went to fdo-packet-m1xl-2
15:52
<
Lyude >
daniels, bentiss - did either of you two get a chance to read the email I sent?
16:00
ezequielg has quit []
16:01
<
bentiss >
Lyude: which mail?
16:01
ezequielg has joined #freedesktop
16:03
<
Lyude >
bentiss: [Input needed ASAP] Things to cover in Gitlab Commit presentation for X.org?
16:04
<
bentiss >
Lyude: I don have that email
16:04
<
Lyude >
bentiss:??? I sent it to site wranglers
16:04
* bentiss
avoided that list :)
16:04
<
bentiss >
more like dodged, not avoided
16:04
<
Lyude >
bentiss: need me to forward it then?
16:04
<
bentiss >
yes please :)
16:05
<
Lyude >
bentiss: sent to your rh email
16:05
<
bentiss >
k, got it now
16:07
<
Yeldham >
Howdy! I would like to ask if the 'shared-mime-info' repo is open for a PR to add MIME types of the Godot engine.
16:25
jarthur has joined #freedesktop
16:33
K`den is now known as Kayden
16:40
rektide has quit [Remote host closed the connection]
17:12
<
Lyude >
Yeldham: this channel is for fdo infra issues, might need to ask somewhere else
17:12
<
Lyude >
(…I thought this used to be in the topic, not sure what happened to it)
17:56
ngcortes has joined #freedesktop
18:13
bengal has quit [Ping timeout: 482 seconds]
18:19
xexaxo has quit [Ping timeout: 480 seconds]
18:41
<
bl4ckb0ne >
freenode to libera migration?
18:43
Kayden has quit [Quit: to lunch and the office]
18:46
<
Yeldham >
Lyude: Any idea as to where I could ask?
19:01
<
imirkin_ >
bl4ckb0ne: we all went to oftc, but the topic went to libera? :)
19:02
<
bl4ckb0ne >
yes oftc, sorry
19:25
<
Yeldham >
Yeah, that seems like the best shot.
19:27
<
daniels >
Lyude: saw it but won’t have much of a chance to do anything with it until Sunday I’m afraid
19:41
Kayden has joined #freedesktop
20:57
Seirdy_ has joined #freedesktop
20:59
Seirdy_ has quit []
20:59
Seirdy has quit [Ping timeout: 480 seconds]
20:59
Seirdy has joined #freedesktop
21:35
danvet has quit [Ping timeout: 480 seconds]
22:06
bengal has joined #freedesktop
22:12
Haaninjo has joined #freedesktop
22:39
bengal has quit [Ping timeout: 480 seconds]
23:16
Kayden has quit [Quit: go home]