@stormi oh I see, I'll check the windows 10 vm's then, thank you

I'm taking another look and it looks like the tools may be under another name... xen-tools I may be going blind at this point if I've missed it this many times.

Ya, for the working machine I just had to run opt/xensource/libexec/xen-cmdline --set-dom0 "xen-pciback.hide=(0000:00:02.0)" and on next reboot I could add the GPU as pass through in XOA. for the other machine this didn’t work but using xcp-center did make it show up as a pass through option

Hello! My apologies -- I hadn't gotten around to trying this solution out until now. I just did and it appears to have fixed the issue! VMs are booting correctly and I can interact with them normally through the console. Can also confirm it works on both the 1.29 Dell FW from my original post, as well as the latest 1.32 FW available as of this post. Much thanks @andyhhp and @stormi for both your time and effort in packaging this -- I greatly appreciate it!

A VM's UEFI variables are not supposed to be erased after a migration . What kind of migration was it? Can you describe the exact operations, and whether the storage is local or shared?

I initially did try grub directly but never got the results I needed. (Could be user error) It's nice to change the resolution down because a lot of our installs have service monitors...ie monitors in the corner that were dusted off and used. I would like to see that "boot to firmware" option either in the boot loader hierarchy or as a select-able toggle. It would help on those few times I need to adjust things but my trigger finger is too atrophied to nail that esc.

@olivierlambert Meant to reply to a message in a different forum - Thanks.

@topsecret Another option is to run a hardware network time server which uses an atomic clock. Then point the NTP client for the systems of your client at that hardware time server. You can choose one of the following hardware time servers:- https://www.microchip.com/en-us/products/clock-and-timing/systems/enterprise-network-time-servers/syncserver-s600 https://timetoolsltd.com/products/network-time-server-appliances/ https://www.oscilloquartz.com/en/products-and-services/ntp-servers https://chronos.uk/products/sync-timing/ https://safran-navigation-timing.com/solution/timing-and-synchronization/ For clients which can't have internet at all even for NTP time synchronisation then running a hardware time server is a must for the network. That way all of the clocks will then be in sync with each other.

There's 2 things: VM placement during VM start, which is decided by XAPI (the host toolstack). Due to the amount of resources, there's no reason to spread the boot on different hosts. This might change depending on various factors. VM balancing while it's running, that's for XO load balancer plugin. But there's also some rules to have enough resources usages to trigger In other words: everything is normal in your case.

@olivierlambert You are totally right. I 'thought' I had added a firewall rule to allow incoming NTP requests, but I added it again and reloaded and now all three hosts are syncing. Thanks!

@melthoab Try changing your static max to 8GB.

See my reply in your other thread. Your VM is either our of disk space or your Redis database is damaged.

@melthoab See the documentation -- https://docs.xcp-ng.org/vms/#️-guest-tools

@KPS it should be "yes" of course. Typo!

@KPS I don't know why my brain has decided to index all the xcp-ng forum, so when I read your post, I immediately remember the post with the clonezilla issue

@hvm said in Guest receiving passthrough SATA controllers does not see attached drives: @Teddy-Astie Hello @Teddy-Astie Have you had the opportunity to review the log file output? The past few weeks I have spent a lot of time trying to find a solution. But unfortunately after all the things I have tried, none of them have solved or did changed something to the problem. We are investigating the possibilities to run our branch locations environments with XCP-NG. This requires a working PoC where passthrough of this SATA controller is necessary.

@Andrew It's interesting because I actually had one of the previous versions of the OVA running in XCP and now that VM halts in the same place so it appears something changed on the XCP-NG side.

@Danp It is quite a varied basket. All VM's are linux. Some VM's migration directly from vmware, some were imported vmdk, some were OVA's. The several Ubuntu VMs used the Ubuntu template, some of the other's were AlmaLinux 8. Most were modified in the Advanced settings after the VM was up and running. All have CPU weight at 256. Hyperthreading and Turbo are all enabled.