I have an array with 2 parity disks and 16 data disks.  I had 1 data disk have read errors and become disabled.  I performed short/long tests on the disk and it was fine so I rebuilt the data disk, moved the server to a new location due to home remodel.  Everything was fine for several days then I had 2 different data disks have read errors and get disabled.  I shut server down thinking cabling may be the issue and I was afraid a 3rd data disk may have read errors and get disabled.

 

Server is still shut down.  I though I would recheck all cabling before restarting and risking a 3rd data disk disable before I can rebuild the 2 data disks that are currently disabled.  I have a LSI controller (flashed to IT mode) and SAS breakout cables.

 

My question is what happens if with 2 parity disks, more than 2 data disks get disabled due to read errors (which might be cabling)?  Since the seemingly only way to correct a disabled disk is to rebuild it, 2 parity disks and 3 or more disabled data disks seems like a data loss?

Did you by any chance grab the diagnostics before shutting down?

 

7 hours ago, moose said:

My question is what happens if with 2 parity disks, more than 2 data disks get disabled due to read errors (which might be cabling)?

 

Unraid doesn't disable more disks then there are parity drives, though it can disable more if there are issues during the rebuild. 

I didn't know Unraid won't disable more disks that there are parity drives.  This would have kept me from panicking and shutting it down.

 

@JorgeB I didn't save diagnostics before I shut down the server.  If Unraid won't disable more disks than there are parity drives, I can boot the server and produce diagnostics.  The server state should be ~ the same as it was when the 2 data disks were disabled due to read errors, correct?

Edited September 10, 20223 yr by moose

17 hours ago, moose said:

I can boot the server and produce diagnostics.  The server state should be ~ the same as it was when the 2 data disks were disabled due to read errors, correct?

You should, the syslog starts over after every reboot, so we can't see what happened, but we can see the SMART report for the disabled disks.

On 9/11/2022 at 4:48 AM, JorgeB said:

You should, the syslog starts over after every reboot, so we can't see what happened, but we can see the SMART report for the disabled disks.

Sorry for the delay in responding.  Attached are diagnostics.  I did not start the array, just booted and generated diagnostics.

(Disk 13 is disabled.  Disk 3 was also having issues and I thought it was also disabled when I shut down the server but now disk 3 shows as ok.)

artoo-detoo-diagnostics-20220912-2030.zip

Edited September 13, 20223 yr by moose

Disk3 looks OK, 13 mostly OK, run an extended SMART test on 13, both show several UDMA CRC errors so it could have been cable related.

Thank you JorgeB.  I ran extended tests on disk 13 again and it started failing, disk 5 also started throwing read errors and was disabled.  Disk 5 also failed extended tests.  I ended up replacing both disk 5 and 13.  The data rebuild completed on the new 5 and 13 disks.  I am up and running with no issues!

 

Thank you for the help!!