/dev/da2 failed to read SMART values - not matching the disk I took out?

[talz13]

I've had an SA120 disk shelf for a few years now, it came with 12x2tb drives, 6 Seagate and 6 Toshiba. I've had 3 of the Seagates fail over the years, and have replaced them with more Toshibas. I'm running a RAID Z3 with all 12 disks (no comment on that setup, it's what I did 4 years ago, and don't have room to move everything off to re-do the array)

Running FreeNAS-11.3-U3.2

Today I got a notice that /dev/da2 failed to read SMART values. So, being the SA120, the disks are labeled down the columns, and across... i.e.:

0 3 6 9
1 4 7 10
2 5 8 11

So, da2 should be the first column, bottom row. And it's matched what I've seen before when other drives failed.

However, both SMART and the Disks screen showed da2 being a SEAGATE ST2000NM0023, when what I ended up pulling out was one of the TOSHIBA MG03SCA200.

After I Offlined and pulled the *Toshiba*, I noticed that there was a bit of a clicking noise coming from what sounds to be da3. Seems to be that is the disk that is actually having problems, but I'm already in the process of resilvering the pool, after swapping out da2.

Looking at all of the SMART info for each of the disks, I can't even find the serial number of the disk I pulled. Is there something I should be doing after disk replacements to update the info? And shouldn't running smartctl on a device be pulling the info from the drive itself? The server has been rebooted several times between the last drive replacements and now.

I'm going to be putting all of the serials in a spreadsheet (layed out in the order the disks are in the bays) after this, but it's still confusing why I'm having this trouble right now.

I have to go out of town tonight, but I'll likely be replacing da3 when I get back, after resilvering is complete.

 

[sretalla]

That assignment is no guarantee to be repeated every time (if you boot with the bottom left drive missing for example, you will get the table below)... don't use those labels to reliably find the same disk, always use serial numbers.

0 2 5 8
1 3 6 9
X 4 7 10

This script from @Spearfoot (get_hdd_temps.sh) will list it out for you...

GitHub - Spearfoot/FreeNAS-scripts: Handy shell scripts for use on FreeNAS servers

Handy shell scripts for use on FreeNAS servers. Contribute to Spearfoot/FreeNAS-scripts development by creating an account on GitHub.

github.com

 

[Yorick]

If you prefer to use the UI, Storage -> Disks lists the name (adaX) and the S/N.

In Storage -> Pools -> Status you can "Edit" a disk via its 3-dot menu and you'll see the S/N as well.

A 12-wide raidz3 seems entirely reasonable to me, particularly for use that's light on IOPS.

 

[talz13]

If you prefer to use the UI, Storage -> Disks lists the name (adaX) and the S/N.

In Storage -> Pools -> Status you can "Edit" a disk via its 3-dot menu and you'll see the S/N as well.

A 12-wide raidz3 seems entirely reasonable to me, particularly for use that's light on IOPS.

So that’s what I was talking about. I looked at that page, and the drive I pulled didn’t seem to match any of the listed serial numbers from that page. Not sure why...

also, if the toshiba drive I pulled is not da2, then it’s probably fine, so I’ll likely swap it in for the disk in the 4th slot that is currently the clicking Seagate drive.

any way to force update the serial numbers and models displayed on that screen if they aren’t reflecting what’s actually in there?