Replacing a "healthy" drive

pdoten

Dabbler
Joined
Mar 27, 2013
Messages
24
A few days ago I heard some off sounds coming from my NAS box which is under my desk. Thankfully not the "click of death" but certainly sounded like it was trying to re-read the same area of a drive. Having all sorts of drives fail over the years I knew this wasn't a good sound. Opening up the console I see a fairly ominous "Device: /dev/ada2, 1 Current unreadable (pending) sectors" message.

I ran a long test of the drive (attached) and sure enough it is on it's way out. I'm not an expert in analyzing SMART results, but I don't see any good news in there other than maybe "(90% remaining). "

Per the instructions here: https://www.ixsystems.com/community/resources/replacing-a-failed-failing-disk.75/
It seems fairly straightforward to replace the disk, mark it offline, powerdown, replace drive physically, power up, select, click replace button, open beer. I am running an older version (FreeNAS-9.3-STABLE-201604150515) but the instructions appear to apply accordingly.

My questions are as follows (and pretty simple really)
1. I am assuming it is still ONLINE because even though it is having some read errors the disk hasn't completely died (yet).
2. If I replace this disk with a larger one, I won't see any increase in capacity until I replace the other two similarly sized drives in the pool? That is my understanding. TIA
 

Attachments

  • HD_2.JPG
    HD_2.JPG
    58.9 KB · Views: 220
  • hd_3.JPG
    hd_3.JPG
    113 KB · Views: 239
  • hd_4.JPG
    hd_4.JPG
    125.9 KB · Views: 244
Joined
Oct 18, 2018
Messages
969
I am assuming it is still ONLINE because even though it is having some read errors the disk hasn't completely died (yet).
Correct, in fact ZFS won't necessarily report a pool as degraded if there exist certain issues with the SMART results. I've replaced and RMA'ed disks which ZFS didn't mark as degraded but that nonetheless has troubling SMART test values.

If I replace this disk with a larger one, I won't see any increase in capacity until I replace the other two similarly sized drives in the pool? That is my understanding.
You are correct. The two general ways to increase the size of a pool is by adding a vdev or growing the size of a single vdev by replacing every drive in that vdev with larger drives.

Also, it is worth repeating that no amount of redundancy is a replacement for a good backup. For example, if during disk replacement you lose another drive in the same RAIDZ1 vdev you'll lose the vdev and thus the entire pool. A good backup is a nice thing to have regardless.
 

pdoten

Dabbler
Joined
Mar 27, 2013
Messages
24
Also, it is worth repeating that no amount of redundancy is a replacement for a good backup. For example, if during disk replacement you lose another drive in the same RAIDZ1 vdev you'll lose the vdev and thus the entire pool. A good backup is a nice thing to have regardless.
I do keep an off-site backup of the most important stuff on optical media. Things like natural disasters and theft are my main drivers of that.
 
Top