Hi,
I’m trying to establish if I did something wrong, or if this is a potential bug.
I manually applied a webmin patch using:
sudo webmin patch Fix to correctly handle standalone `RewriteRule` for SSL redirects work · virtualmin/virtualmin-gpl@5f7f6c2 · GitHub
This triggers systemctl restart webmin
When Webmin stopped, I believe the TAP interfaces that were created by QEMU/KVM processes were destroyed.The TAP interfaces are owned by the QEMU process and are destroyed when the process terminates or when certain system events occur during service restarts.
Then the VMs are no longer reachable and become pretty much useless.
All this takes place without any warning or error.
Could you indicate if there’s something wrong in my configuration?
Restarting webmin should not interfere with the network at all. Are you sure other services were not restarted at the same time?
I’ve seen the tap interfaces stop communicating after the networking service is restarted on RHEL, where NetworkManager is restarted - are you using Netplan?
Something appears to be wrong in my configuration.
I can confirm that restarting webmin kills the cloudmin VMs
The webmin service has KillMode=control-group (the default).
This might cause that restarting webmin, systemd sends SIGTERM (signal 15) to all processes in the webmin cgroup, including any VMs that were spawned by Cloudmin?
Just in case it makes any difference, I want to indicate that I installed Cloudmin (and required dependencies) some time AFTER using this server with Webmin + Virtualmin.
I wonder if this could be the reason for the mess?