Hi
I configured proxmox in a cluster with two nodes. Each node has 8 SSD drives which I configured as zfs at raidz1 level. I have a question about the cache option in the virtual machine. Which do you think is the safest in the event of a power failure?
writeback - it seems to be safe when the barrier option is enabled, but this option is probably unavailable for zfs...
writethrough - this option seems good because of the cache on reads and direct writing.
What is your experience?
I will add that the disks are connected via the backplane to the HBA controller on the PCIe. Supermicro servers.
I configured proxmox in a cluster with two nodes. Each node has 8 SSD drives which I configured as zfs at raidz1 level. I have a question about the cache option in the virtual machine. Which do you think is the safest in the event of a power failure?
writeback - it seems to be safe when the barrier option is enabled, but this option is probably unavailable for zfs...
writethrough - this option seems good because of the cache on reads and direct writing.
What is your experience?
I will add that the disks are connected via the backplane to the HBA controller on the PCIe. Supermicro servers.