Upgraded a disk and got this:

 

Last check completed on Tue 5 Nov 2024 08:53:21 AM MST (today)
 Duration: 1 day, 20 hours, 54 minutes, 50 seconds. Average speed: 123.7 MB/s
 Finding 411584 errors

 

Disk 10 (the disk that was replaced) shows writes but zero reads.  From the shell, all of its values during an 'ls /mnt/disks' shows ???????? symbology.

 

The array was not accessible during the entire rebuild (was getting i/o and permissions errors in Windows).  Still getting i/o errors (unable to write to the array at all from within unraid or Windows).  Also, no surprise I guess, my log is at 100%  

 

Seek error rate on that disk is 12114501

 

Diags attached.

 

I assume I need to swap back to the old disk (still have it) and RMA the new one.... Not sure how to do that.

 

 

 

 

ffs2-diagnostics-20241105-1452.zip

Edited November 5, 20241 yr by tucansam

There were many read errors on parity, you should have aborted the rebuild, looks more like a power/connection issue, but the rebuilt disk will likely have data corruption, do you still have the old disk intact?

Yes I have the original drive, but its smaller (14tb) than the problem replacement (20tb) and I don't know how to revert back.

 

I also have a replacement 20tb disk on the way.

 

I am unsure of my next move.

I would recommend doing a new config with the old drive, check parity is already valid, but then run a correcting parity check.

To be clear, because I have nearly 180TB at risk here....

 

Stop array, new config, unassign Disk 10.  Start array, parity rebuild occurs.  Data on Disk 10 is not available.

 

Once parity is rebuilt, stop array, new config, put old Disk 10 back in, tell unraid to preserve everything, run parity check.  Then data on old Disk 10 will be availale.

 

Is this the correct work flow?

 

Thanks.

No, no rebuilds, install the old disk10, do a new config and assign it in place of the current one, then check parity is already valid, start the array and run a correcting parity check, before this recommend checking/replacing cables for parity.

 

Any doubts ask.

Thanks Jorge.  I replaced the old disk as per above and started a correcting parity check as recommended.  

 

I'm 15 hours and 26 minutes in, at 21.7% done with 1 day and 20 hours remaining, with 6.08 million sync errors corrected so far....

 

Should this be of concern?

Just sync errors are not necessarily a problem, but I would not expect that many unless there were a lot of changes to disk10, post current diags to take a look.

Thanks.

-ffs2-diagnostics-20241111-1049.zip

No disk errors, just sync errors, so just let if finished, and 6 Million is not a lot of sync errors, they are by sector, one single file could do that.

Thank you sir!