Thanks Hector. That drive cannot hit it: `virt_boundary_mask=4095` means the controller has no SGL support, so nvme-pci keeps the 4 KiB virtual boundary, splits at every sub-page gap and never takes the IOVA- coalescing path. The bug needs `virt_boundary_mask=0`, i.e. an SGL- capable (in practice enterprise) NVMe. I left that condition out of the list above:
- `/sys/block/nvme0n1/queue/virt_boundary_mask` must be `0`. Client drives such as the PM9A1 show `4095` and are not affected. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2169551 Title: VMs on ephemeral fails to boot [arm64, amd64] To manage notifications about this bug go to: https://bugs.launchpad.net/charm-nova-compute/+bug/2169551/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
