Hi,

been Unraid user for 10+ years, from version 3 or 4 and now on 6.83.  

One of my disks failed (disk2) and I started a repair from the parity by introducing a brand new disk; nothing new for me, did that in the past.

Now, during the rebuild the other disk #5 started reporting read errors:

 

Nov 15 22:21:08 Tower kernel: print_req_error: I/O error, dev sdg, sector 473325040

Nov 15 22:21:08 Tower kernel: md: disk5 read error, sector=473323952

Nov 15 22:21:08 Tower kernel: md: disk5 read error, sector=473323960

Nov 15 22:21:08 Tower kernel: md: disk5 read error, sector=473323968

Nov 15 22:21:08 Tower kernel: md: disk5 read error, sector=473323976......

 

with the total of 77,056 errors. The rebuild completed and all disks and array is green.

 

Now, I am running a parity check (with NO correction) and getting same read errors:

Nov 16 08:15:47 Tower kernel: md: recovery thread: P incorrect, sector=473323952

Nov 16 08:15:47 Tower kernel: md: recovery thread: P incorrect, sector=473323960

Nov 16 08:15:47 Tower kernel: md: recovery thread: P incorrect, sector=473323968

Nov 16 08:15:47 Tower kernel: md: recovery thread: P incorrect, sector=473323976.....

 

The parity check will complete in next couple of hours however not sure if I want to repeat the parity sync again with the correction or start looking for disk5 replacement?

The smart report for Disk5 is attached.

 

Please advise.

 

regards

Bobtower-smart-20221116-1259.zip

 

 

SMART looks fine, diags might show more, if you didn't reboot yet, also keep in mind that the rebuilt disk will likely have some corrupt data, unless the read errors didn't coincide with used sectors.

Thank you JorgeB,

 

No, I have not rebooted yet.  What would be the next step? 

Your help is appreciated. 

Post the diagnostics.

 

Diagnostics attached as requested

tower-diagnostics-20221116-1329.zip

Looks more like a power/connection problem, if there weren't any writes to the array you can try rebuilding again after checking/replacing the cables on that disk, but only rebuild on top if the emulated disk mounts and contents look correct.

Nothing was written since the rebuild. The rebuilt disk2 looks fine; same with disk5 that have read errors.

I will check the cabling but dont know how to rebuild again. 

Edited November 16, 20223 yr by BobNik
typo

To rebuild again:

-unassign disk2

-start the array, check that the emulated disk2 mounts and contents look correct, if yes

-stop array

-re-assign disk2

-start array to begin rebuild.

 

Thank you.

Will follow your steps later tonight.

Update:

All cables checked and rebuild of disk2 completed with no errors from any other disks. Currently running parity check (with correction on).

Parity completed with 0 errors...everything is back to normal.

Thank you JorgeB for the help.