Hi Rohit, thank you for your reply. We got it working like this indeed. Main Problem was to configure the Hypervisor and Drivers for that but we got that as well. VGPUs work like a charm now.
Thanks! On 2020/12/12 06:59:55, Rohit Yadav <[email protected]> wrote: > Yes, you want to first figure out the domain XML that is necessary to add the > device, and then try to use the extra config feature. See: > https://www.shapeblue.com/cloudstack-feature-first-look-enable-sending-of-arbitrary-configuration-data-to-vms/ > https://cwiki.apache.org/confluence/display/CLOUDSTACK/Enable+additional+configuration+metadata+to+virtual+machines > > > Regards. > > ________________________________ > From: Vincent Hermes <[email protected]> > Sent: Wednesday, December 9, 2020 22:17 > To: [email protected] <[email protected]> > Subject: 4.14 KVM GPU > > Hi Guys, > > we got this constellation: > 1x CS MGMT Server 4.14 > 6x CentOS7 with KVM > 1x Primary Ceph Storage > 1x Secondary NFS Storage > > Now we got two additional CentOS7 KVM Hosts with NVIDIA RTX8000 inside and we > want to passthrough it to VMs on these hosts. > > We got plenty of information from the internet but we dont want to destroy > everything testing all of the possibilities. > > ---------- > Is there a short way to passthrough GPUs with KVM? > ---------- > > Thanks in advance > Vincent > > [email protected] > www.shapeblue.com > 3 London Bridge Street, 3rd floor, News Building, London SE1 9SGUK > @shapeblue > > > >
