New Rust Xen guest tools

yann

@john-c a rolling-release distro does not have to follow every commit of each software component, that would be a bit extreme They are rather about providing users with latest component releases as they are published.

mhomoky

@olivierlambert / @Sam According to the Gitlab docs the project URI needs to be in URL encoded format, resulting in a working line as below in my own cloud-init:

      deb [trusted=yes] https://gitlab.com/api/v4/projects/xen-project\%252Fxen-guest-agent/packages/generic/deb-amd64/ release/

I found this through experimentation, but weirdly it seems to be a double-URL encoded string at that point in the URI for Gitlab… go figure.

I have another escape related question though. I'm not sure if it's your YAML parser or my editor but if I attempt to use the % as found in the beginning of the documentation example, neovim no longer recognises the file as YAML. So this:

#cloud-config
hostname: {name}
fqdn: {name}%.my-domain.com

from line 3 onwards doesn't work correctly, despite being valid YAML 1.2.2.

Again to get this working, I had to surround the value with single quotes:

#cloud-config
hostname: '{name}%'
fqdn: '{name}%.my-domain.com'

Which library / version of YAML are you targeting?

Also, is the UUID or mac address of a VM exposed to the cloud-init parser internally (or is there a list somewhere of other variables like {name} that are?).

Thanks for a great product and I hope this helps someone

olivierlambert

Hi @mhomoky I think it's unrelated to the topic, that would be better to create a new thread I think.

Sam

@mhomoky Thanks, worked perfectly fine.

deb [trusted=yes] https://gitlab.com/api/v4/projects/xen-project\%252Fxen-guest-agent/packages/generic/deb-amd64/ release/

No issues in Cloud-config

Also for hostname, I just used {name}without any %

The only thing missing is being able to reapply cloud-init configs or easy reconfigure from the Ui. But that's for another topic.

dClauzel

@olivierlambert I tried it in Debian 12.8/stable with sysvinit as init (and not systemd).

The package was installed without problem, but the agent is not started.

When run manually via sudo /usr/sbin/xen-guest-agent, the agent is then correctly detected by the dom0.

So it means no dependencies on systemd (which is good :), and only an init script in /etc/init.d/ is required for having full support.

yann

@dClauzel since Debian has switched away from sysvinit we did not take the time to provide an init.d script, but if there is a need, we can include one. Feel free to open an issue (or, even better, a merge request ) on the project in Gitlab!

Thanks for your feedback!

dClauzel

@yann Thanks for the fast reply

Ok, I am opening a ticket, and providing a simple support script.

David_5.1

Hello,

I have a setup with routers (running Debian Bookworm) that need more interfaces than the maximum allowed number of vifs so the VM has one vif and VLAN untagging is done inside it.
As no IPs were detected by the traditional xen tools, I switched to the new, Rust ones, but they fail to detect IPs too

I think it is caused by the lack of /sys/class/net/enX0.*/device/devtype files, despite a few enX0.* interfaces (and folders) and /sys/class/net/enX0/device/devtype (containing vif, as expected) exist.

Should I create an issue directly in the Gitlab project, or comment the partially-related one for SR-IOV support?

yann

@David_5.1 a new issue would be better. Thanks fo your feedback!

David_5.1

@yann here you are : https://gitlab.com/xen-project/xen-guest-agent/-/issues/22

Dillweed

@kevdog @yann said in New Rust Xen guest tools:

@kevdog great news, looking forward for this PKGBUILD!
Wouldn't it make sense to build from release packages rather than from Git?

The CI scripts should give you some guidance. For dependencies you should have a list at https://gitlab.com/xen-project/xen-guest-agent#build-requirements. Not sure why you would want python-setuptools?

Hey all, I put together PKGBUILD for Arch which pulls the latest git. It seems to be working fine for me. You'll need to install xen package on AUR.
https://aur.archlinux.org/packages/xen-guest-agent-git

It's my first attempt at submitting something on AUR. I look forward to any feedback.

Andrew

@olivierlambert No IP record... Using Debian 11 with Management agent 1.0.0-proto-0.4.0. It's a non-standard interface setup with the IPv4/IPv6 assigned to the bridge interface. The agent does not report any addresses up to XO.

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master br0 state UP group default qlen 1000
    link/ether 32:a9:24:28:18:56 brd ff:ff:ff:ff:ff:ff
3: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 52:31:46:59:66:a1 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.33/24 brd 192.168.1.255 scope global br0
       valid_lft forever preferred_lft forever
    inet6 2000::5031:46ff:fe59:66a1/64 scope global dynamic mngtmpaddr
       valid_lft 2591833sec preferred_lft 604633sec
    inet6 fe80::5031:46ff:fe59:66a1/64 scope link
       valid_lft forever preferred_lft forever

olivierlambert

Likely because br0 isn't parsed. Pinging @yann and @TeddyAstie

yann

@olivierlambert yes, and that's a known issue. The protocol used to communicate with XAPI only allows to report info for VIFs (and SR/IOV, with support coming with in a PR). We can likely implement something by querying the status of bridge devices and listening to their changes like we do for the VIFs, and report those for the VIFs that are part of bridges - but it's a bit more than just "parsing br0" .

Opened https://gitlab.com/xen-project/xen-guest-agent/-/issues/24

flakpyro

With the release of Debian 13 apt now complains that the repo is not signed. Also Debain has changed to using .sources files for repos.

For example,, the new format would be:

Types: deb
URIs: https://gitlab.com/api/v4/projects/xen-project%2Fxen-guest-agent/packages/generic/deb-amd64/
Suites: release/
Components: 
Signed-By: https://path/to/release.gpg
Trusted: yes

Maybe worth addding a release.gpg to the repo and updating documentation when configuring the repo on newer Debian / Ubuntu releases?

Example of the error when no key is present:

Ign:8 https://gitlab.com/api/v4/projects/xen-project%2Fxen-guest-agent/packages/generic/deb-amd64 release/ InRelease
Hit:9 https://gitlab.com/api/v4/projects/xen-project%2Fxen-guest-agent/packages/generic/deb-amd64 release/ Release
Ign:10 https://gitlab.com/api/v4/projects/xen-project%2Fxen-guest-agent/packages/generic/deb-amd64 release/ Release.gpg
Fetched 176 kB in 1s (154 kB/s)
All packages are up to date.    
Notice: Missing Signed-By in the sources.list(5) entry for 'https://gitlab.com/api/v4/projects/xen-project%2Fxen-guest-agent/packages/generic/deb-amd64'

olivierlambert

@yann can you update the README accordingly?

yann

@olivierlambert updating the README will be quick enough... but if the sig is indeed mandatory we need to setup something for this first... and autosigning from a CI rather requires doing that on a trusted runner rather than on gitlab-provided ones, so that requires some provisioning and IT work first.

yann

@flakpyro the old format is still supported, and actually the [trusted=yes] in the old-style configuration shown in the release notes does work in my quick test with our own Debian 13 hub template.

David_5.1

One-line format should work fine with Trixie, but as the “new” deb822 format has been supported since Debian Jessie, it should be usable on most installs.

Jessie manpage for reference : https://manpages.debian.org/jessie/apt/sources.list.5.en.html#:~:text=rfc822

john.c

@yann Though the deb822 format allows for that file in sources format, to have the signing key tied to that file’s specified repositories. Very important as it ensures that the key is only used by that repository, unless otherwise specified. The old format typically tends to apply that key to all repositories. So even repositories which shouldn’t use it could, worse the key was trusted for all repositories by the client.

In the new format the repositories can have the specific key tied to them, on the client side as well as the server side.