Hi all,
My current server (24-bay chassis) consists of 4x 6-wide RAIDZ2 vdevs with 10TB drives. I'm sitting at 95% used and one of the drives is faulted. With the sale of 18TB drives, I'm looking to expand, but struggling to land on a new ZFS topology.
I'm not happy with 6-wide RAIDZ2. It wastes too much space and I don't need the performance. I have a pool of mirrored SSDs for VMs/apps/etc. The HDD pool is strictly for Plex, mostly sequential reads and writes.
I have a 45-bay chassis on standby that I can use, and am thinking about eventually upgrading to a 60-bay chassis, so I'm looking for a topology that can fully utilize a 60-bay chassis. I'm not interested in hot-spares.
Initially I was thinking of changing to 3x 15-wide RAIDZ3. I would buy 15x 18TB drives and re-use 15 of the 10TB drives from the old pool and use the rest as cold spares. I can then add another vdev of 15 when I upgrade to a 60-bay chassis.
Alternatively, I could run 3x 10-wide RAIDZ2. I would buy 10x 18TB drives and re-use 20 of the 10TB drives from the old pool. I can then add another two vdevs when I get the 60-bay chassis. This configuration has the same space efficiency (80%), but is easier to upgrade and has better IOPS performance (although that isn't needed). I guess the only drawback is it is slightly less fault tolerant.
Anyone have thoughts on ZFS topology with 60 drives? Thank you for any insight you can provide.
Hardware:
Motherboard: SuperMicro MBD-X11SSH-LN4F-O
CPU: Intel Xeon E3-1230 V6
RAM: 32GB DDR4-2133 ECC
HBAs: 3x LSI 9211-8i flashed to latest IT fw
Case: Norco 4224
Drives: 24xWD100EMAZ
TrueNAS-12.0-U8
My current server (24-bay chassis) consists of 4x 6-wide RAIDZ2 vdevs with 10TB drives. I'm sitting at 95% used and one of the drives is faulted. With the sale of 18TB drives, I'm looking to expand, but struggling to land on a new ZFS topology.
I'm not happy with 6-wide RAIDZ2. It wastes too much space and I don't need the performance. I have a pool of mirrored SSDs for VMs/apps/etc. The HDD pool is strictly for Plex, mostly sequential reads and writes.
I have a 45-bay chassis on standby that I can use, and am thinking about eventually upgrading to a 60-bay chassis, so I'm looking for a topology that can fully utilize a 60-bay chassis. I'm not interested in hot-spares.
Initially I was thinking of changing to 3x 15-wide RAIDZ3. I would buy 15x 18TB drives and re-use 15 of the 10TB drives from the old pool and use the rest as cold spares. I can then add another vdev of 15 when I upgrade to a 60-bay chassis.
Alternatively, I could run 3x 10-wide RAIDZ2. I would buy 10x 18TB drives and re-use 20 of the 10TB drives from the old pool. I can then add another two vdevs when I get the 60-bay chassis. This configuration has the same space efficiency (80%), but is easier to upgrade and has better IOPS performance (although that isn't needed). I guess the only drawback is it is slightly less fault tolerant.
Anyone have thoughts on ZFS topology with 60 drives? Thank you for any insight you can provide.
Hardware:
Motherboard: SuperMicro MBD-X11SSH-LN4F-O
CPU: Intel Xeon E3-1230 V6
RAM: 32GB DDR4-2133 ECC
HBAs: 3x LSI 9211-8i flashed to latest IT fw
Case: Norco 4224
Drives: 24xWD100EMAZ
TrueNAS-12.0-U8