daniels 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
gegoxaren[m] has quit [Quit: Reconnecting]
gegoxaren[m] has joined #freedesktop
privacy has joined #freedesktop
karolherbst has quit [Quit: Konversation terminated!]
karolherbst has joined #freedesktop
MrBonkers has quit [Quit: Ping timeout (120 seconds)]
MrBonkers has joined #freedesktop
MrBonkers has quit []
MrBonkers has joined #freedesktop
<Ford_Prefect>
whot: happy to move elsewhere if it makes sense. The use-cases might be things like pair progamming, collaborative editing, or mirroring for support, where you might want to capture and send the input out over the network, inject remote input, but also let local input be functional locally. Like a multiplayer version of RD, I suppose.
<Ford_Prefect>
(the first of those is what I'm think about rn, but seemed like it could be useful for more)
lsd|2 has joined #freedesktop
<whot>
Ford_Prefect: so, the thing is that it's easy to imagine use-cases but it's best to focus on use-cases that *need* to be fixed fsvo immediately. And often it may be better to add a specific portal for this use-case instead of using other portals that may not match the intended use-case. capturing and re-injecting seems different enough that it probably needs special treatment
<Ford_Prefect>
whot: fair enough, let me try to capture something more concrete and file a bug so it's a better documented discussion. I wasn't sure what was currently possible, so this is already quite helpful.
<whot>
Ford_Prefect: ftr, the original proposal was to have an input emulation portal and then it later got folded into RD instead because it had enough overlap (and exiting integration)
<whot>
the original proposal for libei, I mean
Leopold_ has quit [Remote host closed the connection]
Leopold has joined #freedesktop
<Ford_Prefect>
Ack. Combing the two isn't unreasonable from what I see. I'll also try to look at what Zoom does on other platforms too, to have more concrete use-cases (though I do take your point of solving things that will actually be solved now rather than enabling the universe of possibilities)
<whot>
the fewer xkcd workflows we enable, the better for our long-term sanity :)
MrBonkers has quit [Quit: Ping timeout (120 seconds)]
MrBonkers has joined #freedesktop
bmodem has joined #freedesktop
Mithrandir has quit [Quit: brb]
Mithrandir has joined #freedesktop
Haaninjo has joined #freedesktop
Haaninjo has quit []
psykose has quit [Remote host closed the connection]
psykose has joined #freedesktop
MrBonkers has quit [Quit: Ping timeout (120 seconds)]
MrBonkers has joined #freedesktop
mripard_ has quit []
mripard has joined #freedesktop
ximion has quit [Quit: Detached from the Matrix]
todi has joined #freedesktop
sima has joined #freedesktop
mvlad has joined #freedesktop
AbleBacon has quit [Read error: Connection reset by peer]
<hakzsam>
bentiss: did you see my message about emails not delivered?
<bentiss>
hakzsam: I don't see errors in emails in the past 24h. Maybe because we sent too many of them in a short period after fixing the certificate we are not liked by some ISPs?
<hakzsam>
maybe, I don't know but some are definitely missing on my side
Leopold has quit [Remote host closed the connection]
Leopold has joined #freedesktop
thaller is now known as Guest10163
thaller has joined #freedesktop
Leopold has quit [Remote host closed the connection]
Leopold has joined #freedesktop
Guest10163 has quit [Ping timeout: 480 seconds]
thaytan has quit [Ping timeout: 480 seconds]
i509vcb has quit [Quit: Connection closed for inactivity]
Leopold has quit [Remote host closed the connection]
bl4ckb0ne has quit [Remote host closed the connection]
emersion has quit [Remote host closed the connection]
bl4ckb0ne has joined #freedesktop
emersion has joined #freedesktop
pjakobsson has quit [Remote host closed the connection]
janek has joined #freedesktop
Leopold has joined #freedesktop
DodoGTA has quit [Quit: DodoGTA]
DodoGTA has joined #freedesktop
DodoGTA has quit [Quit: DodoGTA]
thaller is now known as Guest10170
thaller has joined #freedesktop
DodoGTA has joined #freedesktop
Guest10170 has quit [Read error: Connection reset by peer]
pulsar12 has joined #freedesktop
ximion has joined #freedesktop
bmodem has quit [Ping timeout: 480 seconds]
bmodem has joined #freedesktop
ximion has quit [Quit: Detached from the Matrix]
bmodem has quit [Remote host closed the connection]
bmodem has joined #freedesktop
vkareh has joined #freedesktop
pulsar12_ has joined #freedesktop
dcunit3d has quit [Ping timeout: 480 seconds]
bmodem has quit [Ping timeout: 480 seconds]
pulsar12 has quit [Ping timeout: 480 seconds]
pulsar12_ has quit [Read error: Connection reset by peer]
janek has quit [Remote host closed the connection]
konstantin_ has joined #freedesktop
konstantin has quit [Ping timeout: 480 seconds]
vkareh is now known as Guest10202
vkareh has joined #freedesktop
Haaninjo has joined #freedesktop
Guest10202 has quit [Ping timeout: 480 seconds]
shashanks has joined #freedesktop
konstantin_ is now known as konstantin
a-l-e has joined #freedesktop
a-l-e has quit []
mvlad has quit [Remote host closed the connection]
mvlad has joined #freedesktop
scrumplex_ has quit [Quit: Quassel - Signing Off]
scrumplex has joined #freedesktop
thaller is now known as Guest10221
thaller has joined #freedesktop
Guest10221 has quit [Ping timeout: 480 seconds]
alanc has quit [Remote host closed the connection]
alanc has joined #freedesktop
ximion has joined #freedesktop
Leopold has quit [Remote host closed the connection]
<alatiera>
we might have a runner deployed tomorrow, almost done with the hyper-v containers
i509vcb has joined #freedesktop
<jenatali>
alatiera: Thanks!
Leopold has joined #freedesktop
<jenatali>
Just to confirm, that means that we have to switch our containers to use hyper-v isolation instead of process isolation if we want to be able to use that runner?
<alatiera>
jenatali: you hopefully won't need to do anything, it's transparent
<alatiera>
same docker runtime
<jenatali>
Oh ok
<alatiera>
only a switch in the daemon.json for dockerd
<alatiera>
the hyperv containers are a bit more isolated
<jenatali>
I thought it was also part of the dockerfile
<alatiera>
there were 1-2 things I need to fix in gstreamer to get them to work, they actually enforce not touching outside of the CI_PROJECT_DIR for example
<alatiera>
and I guess you might have some issues with 3d devices? although I'd expect that to work as is, given that the runners didn't have a gpu anyway and everything is software rendered when needed
<jenatali>
Oh nevermind, I see what I was thinking of. There's a comment in one of our dockerfiles that indicates that it doesn't work with process isolation :)
<alatiera>
that's impossible
<jenatali>
Yeah, all of our D3D-based tests use WARP, the software rasterizer
<alatiera>
cause our runners were only process isolation so far!
<alatiera>
oh also, hyper-v containers are like 2x slower
<jenatali>
Right, it's just a comment saying that trying to run it with process isolation fails due to networking issues
<jenatali>
I haven't tried to build any of these containers locally, but the guy who wrote that comment apparently did. Clearly it works in CI though. Or at least it did once?
<alatiera>
heh
<jenatali>
I tried to rebuild that container a while ago and it started silently failing so I dunno
V has quit [Ping timeout: 480 seconds]
V has joined #freedesktop
<pinchartl>
is there documentation on how the gitlab.fd.o shared runners are configured ? I'm registering my own runner to spare a download of a 231GB image on the shared runners, and I don't know to select for some of the configuration options. an example that has been tested would be helpful
<jenatali>
alatiera: Btw it seems like we might actually have budget available to spin up an Azure runner on our side. Assuming we're successful, I might be pinging you for help getting it set up - unless there's someone else I should be talking to instead?
<alatiera>
the setup for the windows side is trivial, sure ask
<pinchartl>
alatiera: is the setup for the linux side trivial too ? :-)
<alatiera>
mostly, there's a cloud-init file somewhere
<pinchartl>
I have a feeling that everything will be trivial except the configuration for my precise use case
<alatiera>
it's trivial once you learn everything you need to know!
<pinchartl>
I managed to deploy a gitlab-runner container, and run the register command, so it's "just" about making sure I get the configuration parameters right
<pinchartl>
hehe
<pinchartl>
I've been learning so much the past few weeks. and I got sick today. not sure the two are related
<daniels>
pinchartl: it's less straightforward than you'd hope, I'm sure:)
<daniels>
gitlab-runner-packet.sh generates the configuration and provisions a new machine in Equinix
<pinchartl>
thanks
<pinchartl>
helm
<pinchartl>
another thing to learn
* pinchartl
runs away
<daniels>
dw, it's not helm
<daniels>
so you don't have to worry about helmfile -> helm -> kustomize -> k8s :P
<daniels>
anyway, generate-cloud-init.py creates the cloud-init file which is run once the machine's been provisioned, which sets up the filesystem and all kinds of first-time init
<daniels>
then gitlab-runner-register.sh is what executes gitlab-runner register
<pinchartl>
thanks ! that's very helpful
<pinchartl>
do I understand it right that you install gitlab-runner on the bare metal machine ?
<pinchartl>
(back in my days bare metal meant assembly programming without an OS)
<pinchartl>
(now it means Linux userspace in a VM)
<daniels>
heh :) but *not in a VM?
<pinchartl>
also, while you're here
<pinchartl>
`podman container run --rm` vs. `podman container create ; podman container start`
<pinchartl>
is there one way that is better than the other, does it depend heavily on use cases, or is it a matter of taste ?
<daniels>
you're right though - every runner has its own dedicated bare-metal machine; we run Debian on the host, then that runs gitlab-runner inside a container
<pinchartl>
even `podman generate systemd` supports generating systemd unit files for both option
<daniels>
I honestly have no clue, sorry; I've only ever used the former, and didn't know the latter even existed
<pinchartl>
ok :-)
<pinchartl>
the latter is prone to big "oops" moments if you run `podman container prune` I suppose
i-garrison has quit [Remote host closed the connection]
i-garrison has joined #freedesktop
konstantin_ has joined #freedesktop
konstantin has quit [Ping timeout: 480 seconds]
konstantin has joined #freedesktop
mvlad has quit [Remote host closed the connection]
yann-kaelig has quit [Ping timeout: 480 seconds]
konstantin_ has quit [Ping timeout: 480 seconds]
yann-kaelig has joined #freedesktop
privacy has quit [Remote host closed the connection]
vkareh has quit [Quit: WeeChat 4.1.2]
yann-kaelig has quit [Ping timeout: 480 seconds]
martink has quit [Remote host closed the connection]
martink has joined #freedesktop
yann-kaelig has joined #freedesktop
thaller has quit [Remote host closed the connection]
ykaelig has joined #freedesktop
yann-kaelig has quit [Ping timeout: 480 seconds]
yann-kaelig has joined #freedesktop
ykaelig has quit [Ping timeout: 480 seconds]
ykaelig has joined #freedesktop
yann-kaelig has quit [Ping timeout: 480 seconds]
<bentiss>
daniels, pinchartl: we are not using coreos anymore, because it just broke network eventually without a clear explanation, so I reverted them back to debian, but I'm surprise we still use generate-cloud-init.py
<bentiss>
I thought I folded everything in fdo-infra.py at the root of the repo, but apparently not
<bentiss>
pinchartl: so for now I think generate-cloud-init.py is what was deployed last