I reviewed my NAS today and determined that I’m running on borrowed time.
name | power on hours | Reallocated_Sector_Ct (05) |
---|---|---|
sda | 56501 | 0 |
sdb | 43021 | 0 |
sdc | 56497 | 0 |
sdd | 56501 | 0 |
sde | 71716 | 120 |
sdf | 59382 | 0 |
sdg | 18730 | 0 |
sdh | 70350 | 0 |
sdi | 19449 | 0 |
sdj | 71712 | 8 |
sdk | 44838 | 56 |
sdl | 71715 | 16 |
Do you have it in a drive failure tolerant RAID configuration?
2x6 vdev raidz2. All of the ones with failures are in vdev2 so I just ordered 6x 22tb replacements to not only get rid of the bad drives, but also expand the pool
Wait, you can expand raidz pools?
Yes, but only if you replace all the drives in a vdev with larger capacity drives. After the resilvering the pool will grow to the new capacity
So if you slowly replace the drives with bigger capacities (say 1 replacement a month), you won’t actually gain the extra capacity until you’ve gone through and replaced all the drives (which if you did 1 a month, 6 drives would take 6 months)?
Correct. So long as the resilvering finishes you’re golden.
And you’re on a newer updated version of zfs
Cool, thanks for clearing that up. Im building a home server atm and working through which file system to go with. I have 4x enterprise SSDs that im going to start with, and some of it will be used for NAS storage, among other services.
I know zfs has a lot of fans, but im personally more experienced with btrfs, so trying to decide between the two.
I’ve never done raid with btrfs. I’ve only used it for OS disks, so I couldn’t really say it’d be a good option. I do trust zfs though.
At least they aren’t all the same age, so the chances of multiple failures at the same time is reduced!
True, but resilvering all those drives scares ms.
What does silvering mean?
Basically moving blocks to rebalance/redundant a pool
Zfs term for rebuilding a raid array
Neat. Thank you!
Did you handcraft this table from smart data?
Artisinal handcrafted tables.
Yeah, that was the dull part.