GitHub user rishabhjain1997 added a comment to the discussion: Provisioning the control VM failed in the Kubernetes cluster
**Update — qemu-guest-agent autostart issue persists**
Hi @Pearl1594 and @anirudh09041 — quick update on the `patch.sh` / qga
investigation.
### What we did
1. Disabled the two new hosts (`cloudstack-node2.local`,
`kvm2.cloudstack.local`) so VMs are forced onto `cloudstack.local`.
2. Patched the Ubuntu cloud image (`/var/www/html/ubuntu-cks-fresh-qga.qcow2`)
to install `qemu-guest-agent`, `containerd`, and CKS deps via `virt-customize
--install`.
3. Found that `systemctl enable qemu-guest-agent.service` was a no-op because
the unit shipped by Ubuntu has an **empty `[Install]` section** — it's "static"
and relies on udev/device activation:
```ini
[Unit]
Description=QEMU Guest Agent
BindsTo=dev-virtio\x2dports-org.qemu.guest_agent.0.device
After=dev-virtio\x2dports-org.qemu.guest_agent.0.device
[Service]
ExecStart=-/usr/sbin/qemu-ga
Restart=always
RestartSec=0
[Install]
4. As a workaround, manually created the symlink inside the qcow2:
/etc/systemd/system/multi-user.target.wants/qemu-guest-agent.service
-> /lib/systemd/system/qemu-guest-agent.service
5. Discovered the existing Ubuntu-CKS-Fresh-qga-v2 template was registered
against the wrong URL (pointed at the v1 file). Re-registered as
Ubuntu-CKS-Fresh-qga-v3 against the patched file and provisioned
test-cluster-29.
Current state
- VM i-2-143-VM for test-cluster-29 is on cloudstack.local and has been in
Starting for ~28 minutes.
- patch.sh still times out after 5 min:
WARN ... Process [...] for command [.../patch.sh -n i-2-143-VM ...] timed
out. Output is [].
ERROR ... Passing cmdline failed:timeout
- virsh qemu-agent-command i-2-143-VM '{"execute":"guest-ping"}' →
error: Guest agent is not responding: QEMU guest agent is not connected
Hypothesis
The wants/ symlink alone isn't enough. BindsTo= blocks systemd from starting
qemu-guest-agent.service until the virtio-port device unit
(dev-virtio\x2dports-org.qemu.guest_agent.0.device) activates — and either the
udev rule isn't firing in our boot, or the device unit never resolves.
Checking further on this
GitHub link:
https://github.com/apache/cloudstack/discussions/13056#discussioncomment-16846548
----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]
