I recently purchased a SPARE 10G HD to match the four ZFS drives running in my TrueNAS-13.0-U6.8 system. No problems w the running four, but sooner or later…So, I’m wondering: what is the best way of testing this potential replacement drive in advance of actually having to do so when an installed HD fails? I know that should a drive fail, I’m supposed to be able to pull the bad drive, insert a same-capacity replacement and the RAID should seamlessly (?) incorporate the empty drive into the RAID system and rebuild the redundancy compromised by failure of a single HD. But I’ve never done that, and can imagine things going wrong, even w (hopefully) adequate planning. Suggestions, please?
Have a read here as it covers a number of inquiries. See QA near bottom.
Arguably badblocks -c 2048 -b 4096 -wvs /dev/DRIVEHERE is the most common test due to running 4 different write & read passes across the entire disk, followed by a smart long (you can arguably start with smart long, then badblocks, then another smart long).
It’ll take a week, but at least you’ll be fairly confident it is a working drive.
You’ll also need some familiarity with tmux if you’re running it in TrueNAS, because the command stops when the shell closes. Also careful, it is a destructive test; any data on that drive will be gone - don’t run it on anything active in a pool. Uhh what else?.. Careful & triple check before running it on a system that has any data you want to keep. I like to run it on a separate system from my NAS now-a-days.
Whatever you’d do for a data drive…
The keyword is “burn-in”.
2 Likes
FM,
Thanks. Yes, that’s exactly what I needed to know: what to do BEFORE placing a new HD in the NAS to replace a failing or failed one. Sounds like the main thing is to verify the HD outside of the NAS 1st. I’m familiar w those tests and can run them from my Linux PC.
P
1 Like