This can also happen if you have a hot-pluggable device defined in one of your CT's* which
isn't attached.
For example: I have a USB printer which i always turn off after printing (my energy bill is already high enough
). Udev(?) will then destroy the device (/dev/usb/lp0 in my case) and running #vzlist will throw an error complaining about a non-existed device node which leads to the mentioned error/red alert on the webinterface.
This will probably happen for all hot-pluggable devices which aren't attached but defined in your CT.
This problem can be probably be solved by:
- fixing vzlist (to not output the error or redirect to syslog)
- let the responsable proxmox script check for these errors (and ignore them)
- create the devices dynamically in the CT's
For now i'm checking out the last option and are looking for a way to dynamically run the "vzctl set &VEID --devnode ....." command upon device creation (and "vzctl unset ..." upon device removal). I think this can be handled by writing some custom udev rules.. I'll have to dive into the udev documentation for that....so will be continued.
Regards,
Joost
*by setting DEVNODES= in your CT config or set by the "#vzctl set --devnodes $yourdevice" command