Open travier opened 3 years ago
I have a WIP guide to setup FCOS from the rescue system option. Should I push a stripped down version of that into the documentation or do we want only official support there?
I have a WIP guide to setup FCOS from the rescue system option. Should I push a stripped down version of that into the documentation or do we want only official support there?
I'm not sure how I feel about it TBH. Obviously would love to have official support.
It is fine if we are not confortable with publishing that kind of workaround guide in the official documentation. I can publish it on my blog for now so that it will be picked up as an unofficial solution by search engines.
@travier Are you in touch with the folks working there? What do we need to do to get an FCOS template added in the bare metal workflow?
I don't have any special contact there as I only rent a personal OVHcloud VPS.
I think we need to reach out to them to be put in contact with the technical teams there to figure out how to officially support FCOS. I can try to call them / ask to be called if you think this could help.
I think we need to reach out to them to be put in contact with the technical teams there to figure out how to officially support FCOS. I can try to call them / ask to be called if you think this could help.
I think that'd be cool. Especially since we'd be getting it "for free" since we wouldn't actually need to publish any new artifacts. Let's discuss it in the community meeting tomorrow!
Let's discuss it in the community meeting tomorrow!
Actually, both @travier and I will be on PTO tomorrow. I'll leave it to @travier or whoever will lead the meeting (@dustymabe ?) to determine whether to discuss tomorrow or hold it until next week's meeting.
Let's keep it for tomorrow and I will try to make it. If I don't we'll bump it to next week.
We discussed this in the community meeting today.
11:59:26 dustymabe | #info provided it doesn't take a huge amount of time and
| effort we'd like to work with OVH cloud to get Fedora
| CoreOS as a named image for their various cloud offerings.
| The VMWare offering is particularly interesting because we
| currently don't have any automated test coverage of VMWare.
@travier and @brianredbeard took action items to start investigating OVH cloud a bit more.
I got confirmation that any the BYOI API is a Nova Openstack-like API and thus should most probably work with our OpenStack images.
There's this: https://gist.github.com/felixkrohn/b955c9e691c683b991eb31cd19d22357
Getting a 500 error trying to log in to us.ovhcloud.com, but last I sent a support ticket they said that Fedora CoreOS would be supported but they couldn't give an estimate of when...
iPXE should work too indeed. I did not know this was supported.
Another option is to install via the rescue boot (https://tim.siosm.fr/blog/2021/01/03/fedora-coreos-ovhcloud/) but this is more for one off installs as this does not scale 🙂.
OVH has removed EOL/legacy CoreOS Container Linux and has even updated old-school Fedora versions to the latest, but Fedora CoreOS is still missing. Is anybody actually in contact with them?
I was in contact with them but did not hear anything back for a while. Will ping again.
I created support tickets and again got the response that they can't share a timeline on availability. Do you have a more direct line to be able to confirm that they're actually working on this at all?
Another option is to install via the rescue boot (https://tim.siosm.fr/blog/2021/01/03/fedora-coreos-ovhcloud/) but this is more for one off installs as this does not scale slightly_smiling_face.
I get this as a result:
$ podman run --pull=always --privileged --rm \
> -v /dev:/dev -v /run/udev:/run/udev -v .:/data -w /data \
> quay.io/coreos/coreos-installer:release \
> install /dev/sdb -i config.ign
WARN[0000] Error validating CNI config file /etc/cni/net.d/87-podman-bridge.conflist: [failed to find plugin "bridge" in path [/usr/libexec/cni /usr/lib/cni /usr/local/lib/cni /opt/cni/bin] failed to find plugin "portmap" in path [/usr/libexec/cni /usr/lib/cni /usr/local/lib/cni /opt/cni/bin] failed to find plugin "firewall" in path [/usr/libexec/cni /usr/lib/cni /usr/local/lib/cni /opt/cni/bin] failed to find plugin "tuning" in path [/usr/libexec/cni /usr/lib/cni /usr/local/lib/cni /opt/cni/bin]]
Trying to pull quay.io/coreos/coreos-installer:release...
Getting image source signatures
Copying blob a3ed95caeb02 done
Copying blob c81803563304 done
Copying blob 78ff3cbc7e13 done
Copying blob 3a1b54a4e5cc done
Writing manifest to image destination
Storing signatures
ERRO[0019] CNI network "podman" not found
Error: error configuring network namespace for container 790bd614762dbd16b4da55712192d1b37ea31b15ea8db0bd9cd6230edd944aae: CNI network "podman" not found
Will ping again.
F34 is out now too, any word?
Okay before the step involving podman run
I followed this: https://discuss.hashicorp.com/t/failed-to-find-plugin-bridge-in-path/3095/2
$ curl -L -o cni-plugins.tgz https://github.com/containernetworking/plugins/releases/download/v0.9.1/cni-plugins-linux-amd64-v0.9.1.tgz
$ sudo mkdir -p /opt/cni/bin
$ sudo tar -C /opt/cni/bin -xzf cni-plugins.tgz
Also, I was confused by this:
You can then copy your Ignition config into a file on the serverwith scp
So I just mounted the non-OS partition to a directory I made
$ mkdir /mnt/tmp
$ mount /dev/sdb1 /mnt/tmp
$ cp /mnt/tmp/home/nunya/config.ign .
$ umount /mnt/tmp
then continued at the last step, and it worked!
Edit: on my most recent attempt, I ran out of space on the device but running apt-get clean
allowed me to continue
Edit: Also, there's issues with udevadm
version mismatches that result in the command ending with:
Read disk 619.1 MiB/619.1 MiB (100%)
gpg: Signature made Mon Feb 6 23:11:10 2023 UTC
gpg: using RSA key ACB5EE4E831C74BB7C168D27F55AD3FB5323552A
gpg: checking the trustdb
gpg: marginals needed: 3 completes needed: 1 trust model: pgp
gpg: depth: 0 valid: 4 signed: 0 trust: 0-, 0q, 0n, 0m, 0f, 4u
gpg: Good signature from "Fedora (37) <fedora-37-primary@fedoraproject.org>" [ultimate]
Failed to wait for daemon to reply: Broken pipe
Error: "udevadm" "settle" failed with exit status: 1
Resetting partition table
Error: install failed
So you need to add an option to the final command like -v /usr/bin/udevadm:/usr/bin/udevadm
: https://github.com/coreos/coreos-installer/issues/575#issuecomment-885383071
Folks interested for initial support for this platform in Fedora CoreOS should open an issue with the emerging platform template and follow the steps there. Thanks!
Add support for the OVHcloud platforms.
From Wikipedia:
As there are multiple platforms, this might need to be split into distinct issues. However, as far as I know, the available platforms are based on already supported software platforms which should limit support efforts: