Just added a new Seagate 16TB drive to my Unraid. Initial zeroing took about one day, and went by without a hitch. Formatted the drive, and added it to the array. Then had time to move about 50GB onto it before Unraid warned me about SMART errors. So I cancelled the move and used unbalance to scatter everything off it.

 

SMART overall-health self-assessment test result is PASSED... But here are the counters that errored:

 

  5 Reallocated_Sector_Ct   PO--CK   100   100   010    -    19

187 Reported_Uncorrect      -O--CK   090   090   000    -    10

197 Current_Pending_Sector  -O--C-   100   100   000    -    1
198 Offline_Uncorrectable   ----C-   100   100   000    -    1

 

(full SMART log attached)

 

At this point, the drive is completely empty and has 22 hours of total uptime.. Is it normal to have a few bad sectors, or should I just RMA the drive straight away, as is my immediate inclination? (my first time encountering SMART errors..)

 

Additionally, some questions on how to proceed here:

 

1. I don't have a replacement drive sitting around, so probably looking at 1-2 weeks for a replacement to arrive. In the meantime, can I just remove the faulty data drive from the array (again: it's empty, but not sure if parity "cares") or would I need to fully rebuild parity in that case anyway?
2. Alternatively, could I safely run the array with one missing data drive? Then, once the replacement drive arrives, just put it back in and rebuild? (essentially rebuilding nothing as the drive is empty)
3. Any suggestions on doing a secure wipe before RMA? nwipe?

 

 

fs-smart-20250206-2048.zip

I would RMA it.  In the past I have kept the drive in my machine while they sent me a replacement.  I believe they just took a deposit in case I didn't return the drive once the new one came.  Just explain to them that you cant remove it till you have a physical drive to put in.

3 hours ago, pstmr said:

In the meantime, can I just remove the faulty data drive from the array (again: it's empty, but not sure if parity "cares") or would I need to fully rebuild parity in that case anyway?

You would need to New Config without the disk and rebuild parity. That is what I recommend.

Thanks guys, I'll go ahead and RMA.

 

Just to clarify that I'm understanding this doc correctly: https://docs.unraid.net/legacy/FAQ/shrink-array/

Quote

In general, the best choice is to replace an unwanted or disabled drive by assigning a replacement drive, and allowing Unraid to rebuild it. This way, parity protection for the array is maintained. The following methods are for those who do not want to replace the drive, and are OK with a temporary loss of parity protection.

So basically, the parity rebuild leaves the array with no safety-net for the duration of the rebuild - for both methods. Is that right?

 

 

It just occurred to me that, while I don't have another 16 TB replacement drive, I do have some old 2TB drives kicking around. Could I as a temporary measure just clear, format and rebuild to one of those, and maintain parity while I wait for the postal service to do its thing? Or would that not work and the disk has to be able to account for every bit that was on the original drive, even if only 00000000?

 

Edited February 7, 20251 yr by pstmr
rephrased for clarity

52 minutes ago, pstmr said:

Or would that not work and the disk has to be able to account for every bit that was on the original drive, even if only 00000000?

This, so a replacement drive can never be smaller than the one it replacers (although it can be larger).

On 2/6/2025 at 4:01 PM, pstmr said:

completely empty

4 hours ago, pstmr said:

even if only 00000000?

It's not, an empty disk is not a clear disk.

Thanks again!

I've sent a request for RMA, and following Gragorg's suggestion; asked if they'd send me a new drive immediately with possible deposit, until they receive the old one. If they don't want to do that, I'll go ahead and do new config to rebuild parity.

Edited February 8, 20251 yr by pstmr
can't spell