My Truenas is just up and running for a few weeks and I received an ATA error count from 14 to 29. So I ran a Long Smart test (see uploaded text file for results).
My hardware setup is:
Case: Node 804
Motherboard: Cloud Star MATX-C612
Xeon E5-2680 v4
Memory: 128 GB
Raidz1 Pool 4x 12tb seagate (sda is part of this)
1 cache Vdev 256gig NVME
MirrorPool 2x 8tb for replication
VMDisk 1x 256gig NVME
Version of Truenas:
TrueNAS-SCALE-24.10.2
My feeling tell me the disk is fine and something different causes this warning. But I could be wrong so hopefully someone can analyze the smart test for me. Or tell what to do next.
You could be a victim, run the command smartctl -l farm /dev/sda and look for the FARM data Power On Hours value. It should match or be very close to the SMART Power on Hours.
These are sectors which cannot be written to reliably and have been permanently set aside following repeated failure. Number should ideally be zero. An increasing count literally means that platters are shedding little flakes.
Your drive has been shedding an average of three sectors per hour throughout its short active life…
Spindle Power on Hours: 515214578426
Head Flight Hours: 515214578426
That makes this drive 58,814,449 and a half years old - clearly a Seagate 12TB Stegosaurus Pro (it is 7200RPM after all)!!!
More seriously, the 1,888 reallocated sectors, the 7,186 unrecoverable read errors and the 29 uncorrectable errors are all indicators that you should RMA this drive.
Yes, many of those numbers look fishy to me. The LBAs Read/Written don’t line up, the data is all over the place. Maybe this is an altered drive, I think I said the other day that it wouldn’t take long to figure out how to hack it.
Check the online warranty for the serial number, if it is good, then you should be good.
Update, the seller asked me to check the disks with HD Tune Pro (yep a Windows tool).
Eventually 3 out of 4 gave a warning so I want to change all for 4 disks for a new set. From this point on what would you guys advice me? Backup everything create new pool and write all back? Or is there a smarter/faster way?
My Pool is holding 4 disks in a raidz1 vdev
I’m going to change the 4 disks, whats the easiest wat to do that?
Resilver each one by one? Backup data create new pool of diks? Or do you have other suggestions?
Replace drives to be RMAd one by one with the qualified drives, resilver each time, until the VDEV is healed again.
Scrub pool to eliminate errors
Return drives to be RMA’d
I’d consider looking into goharddrive.com. The used drives they sell are sold with up to a 5 year warranty that I have occasionally used and had zero issues with. Easier than OEM. I highly recommend the HE series from HGST since they use less power than air filled drives.