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
Haaninjo has quit [Quit: Ex-Chat]
jarthur has quit [Ping timeout: 480 seconds]
jarthur has joined #freedesktop
ximion1 has quit []
ppascher has quit [Ping timeout: 480 seconds]
danvet has joined #freedesktop
pinkflames[m] has joined #freedesktop
ppascher has joined #freedesktop
Lyude has quit [Quit: WeeChat 3.2]
alanc has quit [Remote host closed the connection]
<__tim>
one job is downloading at a couple of MB/sec, and other connections are puttering about at kB/sec
<__tim>
but in any case, not sure what changed recently
<ndufresne>
I was testing on artifact.gst and got decent clone speed there
<ndufresne>
(could probably take this opportunity to do a sanity update
<__tim>
git clone on htz3 takes about a minute for me, 4-4.5 MB/sec
<__tim>
(via https)
<ndufresne>
actually, it got stuck at Filtering content: 86% (1374/1583), 260.67 MiB | 79.00 KiB/s
<ndufresne>
system is 99% idle ...
<__tim>
that's the media download, isn't it
<ndufresne>
yes
<ndufresne>
is 147.75.198.156 a known IP ?
<__tim>
isn't that gitlab.freedesktop.org?
<ndufresne>
yep, so it seems stuck on some op toward gitlab
<ndufresne>
during LFS filtering, didn't know that was happening "online"
<daniels>
__tim: I saw it last night on both Wayland/Mesa pipelines, and only htz
<daniels>
so FWIW we don't have per-IP/AS rate limiting or anything; any slowness is a result of weird intermediate networks rather than us being too aggressive upstream
jstein has joined #freedesktop
<ndufresne>
__tim: the artifact server got located in helsinki, perhaps the location matters ?
<__tim>
that routing issue I see with both european locations
<__tim>
but again, that stuff has been around for eternities, so I'm not convinced that's really the underlying cause of our problem
<daniels>
ndufresne: yeah, it occurs with Mesa/Wayland routing to Packet in east-coast US as well, so going to say it's not a Finland-specific issue :)
<__tim>
I think jordan updated the gitlab-runner version a few days ago, not sure if it could be related to that
<__tim>
when did we start noticing issues?
<alatiera>
yea indeed I updated them on monday/tuesday
<alatiera>
weird thing is that its the windows runner too
<__tim>
heh, from their new ashford, virginia location I can git clone at 36MB/sec+
<__tim>
wonder if we can route the .de/.fi htz servers through that, that avoids the other route with the massive packet loss
<__tim>
until they add dedicated servers at that location
<daniels>
hmm, we might be able to give you WireGuard peering into us as well