@pwason said in Using iGPU in VM(s) -- Possible? How?: FirePro S7150 Note: I never had the opportunity to work with AMD vGPU compatible stuff, so I have no idea how this could/should work.

@xisco is this XCP thing? or citrix? about pausing VM - this is default behavior in Hyper-V server and is work flawless - restarting server prior to some MS patching - all vm are going nicely to sleep and after boot resumes accordingly.

Alright. Thanks for letting us know.

I did try that, but it still did not want to recognize and load the PV drivers properly. The VM is still limping with the realtek divers and. It is performing noticeably slow. Oddly others upgraded without a hitch.

smartmontools is available in latest CentOS repositories so it will likely be updated in next release of XCP-ng which will use updated CentOS packages. It's also available in our repository for testing updated CentOS packages with XCP-ng: https://xcp-ng.org/forum/topic/726/experimental-update-of-centos-packages If you don't want to take the risky route, just add the test repository by downloading the .repo file, then edit it to disable xcp-ng-centos_testing by default, and then use it only for installing the updated smartmontools with: yum update smartmontools --enablerepo='xcp-ng-centos_testing' nvme-cli is not available in that repository, but it is from CentOS: https://centos.pkgs.org/7/centos-updates-x86_64/nvme-cli-1.6-4.el7_6.x86_64.rpm.html and I added it to the wishlist at https://github.com/xcp-ng/xcp/issues/56 olivierlambert created this issue in xcp-ng/xcp open List of useful extra packages for dom0 #56

Yes I gave a look but Rancher is so much more different than a regular distribution, even CoreOS. The problem is everything on system is container, I can't install something by default even manually (using the tar.gz) because when you are login, you are in a BusyBox container in some way... IMHO, I don't think the Citrix tools has been design to run in a container (even privileged) and to retrieve information for the entire system that is also in a container... What a mess right ? But in the end, these kind of Docker-ready OSes start giving me headache, even CoreOS now is not working like it was in the past, the fault of a very fast-paced developement from the CoreOS devs, the outdated documentation since they has been buyed by Red Hat (they move to Fedora CoreOS with lot of changes), and Citrix who seems to not giving much interest in CoreOS support.

Perfect ! I was thinking that is only available during the deployment of a VM (I need more experience on Ansible ) I will give a try as soon as I get the possibility to use Ansible.

@atmurphy, I haven't tried it, but there are a set of plugins available for Xencenter that might work under XCP-ng Center that adds an auto-start option. See https://www.citrix.com/blogs/2018/03/14/xenserver-plugins-collection-v6/

@KapilBandlish I don't understand. You are using XCP-ng and talking about Citrix support recommending something? As @borzel said, thin pro on block device via GFS2 is NOT support in XCP-ng, because the storage drivers aren't open source.

High Olivier Number Start (sector) End (sector) Size Code Name 1 46139392 83888127 18.0 GiB 0700 2 8390656 46139391 18.0 GiB 0700 3 87033856 3907028830 1.8 TiB 8E00 4 83888128 84936703 512.0 MiB EF02 5 2048 8390655 4.0 GiB 0700 6 84936704 87033855 1024.0 MiB 8200 Thats the partition tabel at my simpel testing PC 4 VM 1, 2, 5 using 0700 thats for windows NTFS Yesterday i fitting with a reinstall of one 7.5 with RAID1 and a exta LSI RAID Controller. One HDD was damaged, but all my VM was moved to a reseve Sever. So the easyest way was 2 install again. But the install scripts find sometimes one RAID 1 and the Disk behind the LSI RAID. It dont use the /dev/md127 and try to build a new RAID 1 with /dev/sdc and /dev/md127..... and sme craesy options too. Now i am erasing everything on the old HDD, now its running. The Problem with GDISK HDD are one the end of the HDD is a reseve partition tabelle. Most tools dont erase them. But it is very old and cytrix dont fix it. But its bad when you had fix somthing with booting teh hardware server with a live linux

@sfx please take a look here, there are several discussions there: https://github.com/xcp-ng/xcp/issues/107 olivierlambert created this issue in xcp-ng/xcp open Software RAID install on previous used mdadm disks #107

That's mysterious Could we take a look at what you are doing? Could be interesting for the community!

This all seems very over the top for what you're trying to achieve - since there are much simpler solutions out there. BTW, a much better approach than Samsung Dex was Microsoft's Continuum. It, unfortunately, did not catch on in the consumer space (nor will Dex); and, thus, has been relegated to niche business use. I see where you're going with the end concept; but, it is highly unlikely (if not impossible) that such an approach would ever be broadly accepted and implemented in any typical user environment - it's essentially the same management/implementation problem as existing solutions - you're just moving the burden to another portion of the process with no real advantage.

Wooohooooo !! We have a connection [image: 1543525293650-4649e57e-4925-46e0-8aa6-84789150df8c-grafik-resized.png]

@dvdhngs install onto (mdadm/ZFS) RAID 10: no to use (mdadm/ZFS) RAID 10 as SR: yes, of course!

Hi, We are working on mirrors, expect some news in the next weeks

At the moment https://github.com/xcp-ng-rpms only contains repositories for the packages that we had to modify. For other packages, you can find their sources in the source RPMs in https://updates.xcp-ng.org/7/7.6/base/Source/SPackages/