07:05daniels: just use rnm?
07:49bentiss: FWIW, nobody requested more artifacts from the previous servers at Equinix, so I'll return them today or tomorrow
08:18bentiss: and I also just fixed the large file issues with fastly, so I'm re-enabling fastly for gitlab.fd.o now (still with a short TTL in case something else is off)
09:31bentiss: hmm. we seem to still have some first byte timeout went sending artifacts...
09:37bentiss: disabling again fastly, this creates way too much job failures :(
09:54bentiss: fixed in fastly to map the same settings we have in Hetzner, so re-enabling fastly
10:22vjaquez: hi! I'd like to ssh to gstreamer.freedesktop.org (I could access to annarchy in the past) but my keys aren't working anymore. Did something change in that regard? Also, I wonder if I'm using the correct ones... what are the ones for me (vjaquez) in the authorized_keys?
10:38mupuf: bentiss: well, look at you being ready weeks before the deadline
10:43bentiss: mupuf: that was the plan: giving us enough room if it went sideways :)
10:46mupuf: yes, and you did it brilliantly!
11:39bentiss: heh, actually, one missing bit still running on the old server: the maintenance page itself :)
12:49martink: heya! anybody else having issue with fetches and pushes to gitlab.fdo?
12:50bentiss: martink: see https://gitlab.freedesktop.org/freedesktop/freedesktop/-/issues/2076#note_2831847, the part with ssh.gitlab.freedesktop.org
12:57martink: bentiss: talking to ssh.gitlab.freedesktop.org does not help either. I'm trying to determing if it's a local problem I'm experiencing here. if nobody else is seeing issues with ssh.gitlab.fdo then it probably is
12:57bentiss: works fine here
12:57martink: bentiss: thank for the confirmation
12:59__tim: I've seen people report some issues with http git fetches, but they seem to have resolved themselves
13:00__tim: (was sth like `warning: spurious network error (3 tries remaining): error parsing REF pkt-line; class=Net (12)`)
13:01bentiss: __tim: yeah http git fetches is not ideal, as Fastly sometimes closes the connection when we do a POST, and I couldn't see a parameter I could tweak for that
13:01bentiss: ssh is directly talking to Hetzner, so this should be prefered
13:02bentiss: and FWIW, --depth 10 might help when --depth 100 is not working (saw that on the linux-media tree)
13:02__tim: right, but this is for git repo URLs in e.g. subprojects or Cargo.toml where it should just work out of the box behind a corporate firewall etc
13:03bentiss: hmm, maybe this was related to first byte timeouts that I fixed then
13:03bentiss: it used to be too small for our servers
13:04__tim: will let you know if it happens again, last I heard it went away, but might have been luck of course
13:04bentiss: thanks
15:11slomo: bentiss: https://gitlab.freedesktop.org/gstreamer/cerbero/-/jobs/74223172#L785 any ideas what that is about? also happened some days ago when you enabled fastly
15:16bentiss: slomo: the weird part is that the retry worked
15:17slomo: yeah
15:58bentiss: slomo: reproduced locally, and "fixed" by setting CARGO_NET_GIT_FETCH_WITH_CLI=true as told by https://gitlab.freedesktop.org/gstreamer/cerbero/-/jobs/74223172#L802
15:58bentiss: unclear why this isn't working with the internal git lib :/
17:14slomo: bentiss: oh no. we needed that some years ago too, then for a long time it worked. and now it fails again :)
17:14slomo: bentiss: by fixed you mean locally, and I need to re-add that to the ci environment?
17:56robclark: did something recently break w/ fetching via gitlab https url (unrelated to the needed ssh changes)? It was working a few hrs ago but now `fatal: couldn't find remote ref main`
17:58nirbheek_: What does exit code 143 mean for a job? Seems to have failed while uploading artifacts: https://gitlab.freedesktop.org/amyspark/cerbero/-/jobs/74230124
17:58nirbheek_: I thought the hetzner s3 issues were resolved?
18:04robclark: re: fetch errors, this seems to be only about mesa? I can fetch other gitlab trees
18:15daniels: nirbheek_: that's SIGTERM, but ouch, 206k files ...
18:27slomo: bentiss: answering myself, yes need to do that myself. and apparently we did that already for linux, just not for windows :)
18:44__tim: nirbheek_, I think it's a runner issue, it happens on "Cleaning up project directory and file based variables" afaik, and only on the new runners
18:45__tim: new(ish) gstreamer runners that is
18:50nirbheek_: Ah I see
18:50nirbheek_: That should definitely not be an error that fails the job ;)
21:08Lyude: is fdo sick right now? :C
21:08Lyude: can't seem to push the board meeting minutes
21:16alanc: if you're pushing over ssh, make sure you updated the hostname to ssh.gitlab.freedesktop.org
21:17alanc: https://gitlab.freedesktop.org/freedesktop/freedesktop/-/issues/2076#note_2831847 shows how to fix your .ssh/config to do it globally instead of changing every repo
21:21Lyude: oooo ok
21:22Lyude: thank you!