Hello,
I've added a new zVol-backed virtual disk to my VM for bulk storage. The backing storage is a a SATA HDD mirror pool with 8 HDDs in 4 mirrors. The virtual disk is in the RAW format, since it's being stored directly on my PVE node's local ZFS pool.
I've been reading some new guides lately, trying to learn more about reasonable optimizations for data storage when working inside VMs, and I came across this:
Apparently, this disables the ext4 journaling features for the drive, so we can "rely on ZFS" on the underlying real storage.
I generally understand the idea of not needing to use every protective feature of ext4 designed for data integrity on a real physical disk on a virtual disk that's really just a zVol in a ZFS dataset on a mirror pool, but aside from that, what are the real-world implications of this? I assume there's less overhead in general, but is there anything else I should be aware of when using this in production?
Assume I have backups.
Thanks!
I've added a new zVol-backed virtual disk to my VM for bulk storage. The backing storage is a a SATA HDD mirror pool with 8 HDDs in 4 mirrors. The virtual disk is in the RAW format, since it's being stored directly on my PVE node's local ZFS pool.
I've been reading some new guides lately, trying to learn more about reasonable optimizations for data storage when working inside VMs, and I came across this:
tune2fs -O ^has_journal /dev/{nvme dev name}Apparently, this disables the ext4 journaling features for the drive, so we can "rely on ZFS" on the underlying real storage.
I generally understand the idea of not needing to use every protective feature of ext4 designed for data integrity on a real physical disk on a virtual disk that's really just a zVol in a ZFS dataset on a mirror pool, but aside from that, what are the real-world implications of this? I assume there's less overhead in general, but is there anything else I should be aware of when using this in production?
Assume I have backups.
Thanks!
Last edited: