We currently support SCSI Hostdev Passthrough using a scsi-generic driver only <hostdev mode='subsystem' type='scsi' managed='no' rawio='yes'> which doesn't perform that well. We should add scsi-block and scsi-hd and ideally also virtio-blk-pci. - scsi-block has less overhead than scsi-generic with same capabilities. <disk type='block' device='lun' rawio='yes'> we should consider it a default for disk devices and leave scsi-generic for tapes and such. - scsi-hd is a non-passthrough way, suitable for SSDs with higher performance <disk type='block' device='disk'> can use aio-native - virtio-blk-pci would be the most performant way without the SCSI overhead <disk type='block' device='disk'> with <target dev='vda' bus='virtio'/> supports serial number - for completion, there's also the existing PCI Passthrough option suitable for PCI NVMe SSDs As a rule of thumb the performance should follow: SCSI passthrough overhead guest host ---------------------------------------------------------------------- scsi-generic yes higher high scsi-block yes higher low scsi-hd no higher low virtio-blk-pci no lower lower PCI passthrough N/A lower lowest Possible options depend on the kind of the device: - SCSI non-disks can be exposed as passthrough only(scsi-generic) - SCSI disks can be exposed as passthrough(scsi-block since it's superior to scsi-generic and has no disadvantage), as emulated SCSI(scsi-hd), or even as virtio-blk-pci for better performance - PCI non-NVMe devices can be exposed as passthrough only - PCI NVMe devices can be exposed as PCI passthrough, emulated SCSI(scsi-hd) or virtio-blk-pci. They all should honor the iothread setting as well
I agree with the plan, it's a good summary of past discussions and BZs.
WARN: Bug status (ON_QA) wasn't changed but the folowing should be fixed: [Found non-acked flags: '{}', ] For more info please contact: rhv-devops: Bug status (ON_QA) wasn't changed but the folowing should be fixed: [Found non-acked flags: '{}', ] For more info please contact: rhv-devops
Reassigned, see https://bugzilla.redhat.com/show_bug.cgi?id=1793550
Verified: ovirt-engine-4.4.1.7-0.3.el8ev libvirt-daemon-6.0.0-25.module+el8.2.1+7154+47ffd890.x86_64 vdsm-4.40.22-1.el8ev.x86_64 qemu-kvm-4.2.0-28.module+el8.2.1+7211+16dfe810.x86_64 Verification scenario: Polarion test plan added to external trackers.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Important: RHV Manager (ovirt-engine) 4.4 security, bug fix, and enhancement update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2020:3247