Disk Read Errors after power cut; disk OK, or needs replacing?

[PitRejection2359]

Hi - I had a couple of power cuts around a week ago. Unraid has done 2 full parity checks since it came online and has both times had 379 errors show:

 

Disk 3 is showing 256 read errors:

 

Log from shortly after the power came back on:

But SMART check and diagnostics don't really show anything too bad as far as I can see (attached).

 

SAMSUNG_HD204UI_S2H7J90B622655-20230422-1612.txt

tatooine-diagnostics-20230422-1635.zip

 

Is the drive dying and needs replacing, or will a reboot reset the daily fault email and clear the read errors? Thanks in advance,

[JorgeB]

Last check was correct so the sync errors should be fixed, as for the disk, it's logged as a disk problem so run en extended SMART test.

[PitRejection2359]

Yeah, did the extended test, it came back with no errors, as far as I can tell. I thought I had attached it to the first post... Will check shortly.

 

I'm a little confused by the parity check results (my lack of understanding)! The status is "OK", but it shows 379 errors - exactly the same as the previous parity check after the power failure. How can the parity be OK if there is a read error?

[PitRejection2359]

SMART page for the drive:

 

SMART extended test results:

SAMSUNG_HD204UI_S2H7J90B622655-20230501-1329.txt

[JorgeB]

15 minutes ago, PitRejection2359 said:

I'm a little confused by the parity check results (my lack of understanding)! The status is "OK", but it shows 379 errors - exactly the same as the previous parity check after the power failure. How can the parity be OK if there is a read error?

Diags don't show the first one but I assume it was no correct, so it's normal for the 2nd one to find the same errors, when there's a read error Unraid will write that sector back to disk using parity plus the remaining disks, if there's a write error the disk will be disabled, if there isn't the data will be correct, at least as reported by the disk.

[JorgeB]

As for the disk since it passed the extended test is OK for now, replace/swap cables with a different disk to rule that out, any more read errors after that I would replace it, a lot of hours and SMART doesn't look that great.

[PitRejection2359]

 

33 minutes ago, JorgeB said:

As for the disk since it passed the extended test is OK for now, replace/swap cables with a different disk to rule that out, any more read errors after that I would replace it, a lot of hours and SMART doesn't look that great.

Thanks very much, will take a look at the cables and keep an eye on the disk. 👍

[PitRejection2359]

Just thought of another thing - how do I cancel the error that the daily status email reports:

Quote

 

Event: Unraid Status

Subject: Notice [TATOOINE] - array health report [FAIL]

Description: Array has 9 disks (including parity & cache)

Importance: warning

 

Parity - WDC_WD40EZRZ-22GXCB0_WD-WCC7K2FDR7FK (sdf) - standby [OK] Disk 1 - WDC_WD20EURX-63T0FY0_WD-WCC4M5ZS0CA2 (sdg) - standby [OK] Disk 2 - SAMSUNG_HD204UI_S2H7J90B622647 (sdl) - standby [OK] Disk 3 - SAMSUNG_HD204UI_S2H7J90B622655 (sdh) - standby (disk has read errors) [NOK] Disk 4 - WDC_WD30EZRX-00D8PB0_WD-WCC4N2FH6VH5 (sde) - standby [OK] Cache - CT500MX500SSD1_1911E1F21FD5 (sdi) - active 40 C [OK] Cache 2 - SATA_SSD_B7E0079C1B3500486377 (sdd) - active 33 C [OK] Cache 3 - WDC_WDS500G2B0B-00YS70_20283U479210 (sdc) - active 32 C [OK] Cache 4 - CT240BX300SSD1_1745E1064359 (sdj) - active 38 C [OK]

 

Parity is valid

Last checked on Thu 20 Apr 2023 07:48:46 AM BST (10 days ago), finding 379 errors.

Duration: 10 hours, 49 minutes, 27 seconds. Average speed: 102.7 MB/s

 

I've acknowledged the UDMA error which was there after the restart, and tried to "clear stats" from the main page, but can't get rid of those 256 read errors on Disk 3. Do I need to restart the server for that?

[JorgeB]

Try re-starting the array, IIRC that will do it, if it doesn't reboot.

[PitRejection2359]

On 5/1/2023 at 6:04 PM, JorgeB said:

Try re-starting the array, IIRC that will do it, if it doesn't reboot.

Restarting the array didn't reset the read errors, but restarting the server did. Thank you for all your help. 👍