When I reboot the host, I have to use the Force Restart Host feature due to the "no hosts available" error that happens when just using the Restart Host function.

So I used the Force Restart, and what I observed is :

Linux vms all shutdown cleanly (status turns orange)
Windows vm's appear to have been forced shutdown (no orange status).
XOA vm shutdown (clean)
Host restarted.

This host took probably < than a minute to shutdown (amd gen 3 all nvme, it's fast).

All the vms have the xcp-ng client tools installed. All patches installed on host. If I manually shutdown a windows vm from XO then it does shutdown cleanly.

Using XO from source (2dbe35a31c4742d0ee79e40eef43e53bcef4615f) built 12 march. Looking at the commit history I don't see any changes around this area.

I have rebuilt XO from the latest commit today, but cannot restart the host again at the moment - and I have concerns about screwing up the windows vms doing this too often (I do have backups but still) - windows is fragile!

Before I create an issue on github, is this as designed or a bug? My expectation would be that all vms with the client tools installed would be shutdown cleanly.

Is there something I can look for in the log files?