It'd be useful for (more advanced) users to have a page useful for fixing/debugging occasional VM startup failures. Typically these are caused by:
Memory shortage, usually memory being taken away from the VM at some crucial moment during its boot time, which brings the OOM killer into action, and the end result is VM not booting successfully,
Incorrectly assigned devices for some service VMs. These problems might be causes e.g. by:
users erroneously assigning the same device(s) to 2 or more VMs (which is perfectly ok) and then try to start both (which is not, of course),
users restoring VMs (with devices assigned) from a backup, which was made on a different system (which used different BDF addresses, e.g. for the WiFi card),
very common on recent Intel chips: users assigning only one of the USB controllers, while the other leaving in another VM, but due to some god-knows-what-magic both of these devices are actually the same one device, and - as a result of this - Xen being unable to perform proper PCI device reset on the device to be assigned. A workaround for this is to do: qvm-prefs <vmname> -s pci_strictreset false :)
For HVM-based VMs, and specifically for Windows AppVMs: by Windows performing some sort of cleanup (Start in Safe Mode screen, etc). The work around for this is to enable Run in debug mode in the Manager.
It'd be useful for (more advanced) users to have a page useful for fixing/debugging occasional VM startup failures. Typically these are caused by:
Memory shortage, usually memory being taken away from the VM at some crucial moment during its boot time, which brings the OOM killer into action, and the end result is VM not booting successfully,
Incorrectly assigned devices for some service VMs. These problems might be causes e.g. by:
qvm-prefs <vmname> -s pci_strictreset false
:)For HVM-based VMs, and specifically for Windows AppVMs: by Windows performing some sort of cleanup (Start in Safe Mode screen, etc). The work around for this is to enable
Run in debug mode
in the Manager.