Hi, We are executing hundreds of VMs in our hosts (dual AMD EPYC 7002 CPUs + 2TB memory + 8 NVMe disks). To cope with disk failures, we tried different Linux software RAID mechanisms. At high disk usage, all of them faced unexpectedly high CPU load, mostly in the kernel. Therefore we decided not to use software RAID.
As an alternative, can we use blockdev-mirror for keeping a live mirror of each image? The problem seems to be that when Qemu starts and the mirror exists, it does not need to write the whole disk to the mirror again; writing just the updates would be enough. Looking at the code, the incremental sync mode (which probably does what is needed) is not supported. Is using blockdev-mirror a good idea? Is there any reason that blockdev-mirror's incremental sync mode is missing? As the mirror device is updated in Qemu itself, does using virtio devices help? Thank you very much for your time. Best wishes, Ali
