My suggestion is to try 16.10, 17.04 etc to find out which userspace regressed.
-- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1841098 Title: PCI-passthrough instance creation problem RTX2080Ti Status in linux package in Ubuntu: Incomplete Bug description: I am running Openstack Ansible Rocky 18.1.8 on Ubuntu 18.04 nodes. I use PCI-passthrough (IOMMU vfio) on GTX1080Tis and and RTX2080Tis. The 1080s work fine, but 2080 instances take anywhere from 3 minutes to over 3 hours to start with qemu/KVM and the amount of time varies randomly, with the averaging being around 1.5hrs. I tried upgrading the kernel to 5.0.0 and it did not fix the issue. Downgrading the GPU hosts to Ubuntu 16.04 fixed the issue, they now start in 30 seconds or less. Sorry, but I do not have any logs. I couldn't find any errors in any system logs including enabling debugging on libvirtd. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1841098/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
