Find faulty SSD in RAIDZ

gabrioth

Well-Known Member
Oct 4, 2019
43
9
48
42
A few weeks ago zpool status spouted that one of the SSDs in my pool "ssdpool" had experienced I/O errors and that the scrub had had to resilver the pool with parity data from the other two drives.

I moved tha VMs that inhabited that pool to another zpool to avoid data loss.

I'd like to replace the drive, but, for some reason a reboot of the node or similar reset the error message. I did not clear the pool manually.
They are all of the same model and series of Samsung SATA-drive with very similar serial numbers, so now I don't know which drive is the faulty one.

I have tried writing the pool full of media data and run a scrub, no result, so I'm guessing the errors only will happen when some VMs write continous change to the data.

Is there any way to get log data that tells me which drive had the errors? SMART tells me basically the same thing for all three drives...
 
A few weeks ago zpool status spouted that one of the SSDs in my pool "ssdpool" had experienced I/O errors and that the scrub had had to resilver the pool with parity data from the other two drives.

I moved tha VMs that inhabited that pool to another zpool to avoid data loss.

I'd like to replace the drive, but, for some reason a reboot of the node or similar reset the error message. I did not clear the pool manually.
They are all of the same model and series of Samsung SATA-drive with very similar serial numbers, so now I don't know which drive is the faulty one.

I have tried writing the pool full of media data and run a scrub, no result, so I'm guessing the errors only will happen when some VMs write continous change to the data.

Is there any way to get log data that tells me which drive had the errors? SMART tells me basically the same thing for all three drives...
Decided to look at the SMART data again. Turns out I missed a value of "2" in the "ECC_Error_Rate" row. THe other two drives were at 0.