Disk errors - disabled and blk_error

[paulmorabi]

Hi,

 

I've had a perfectly running Unraid system for a long time now and for the first time yesterday, one of the drivers in my array was disabled. The errors in syslog were:

 

Quote

 kernel: blk_update_request: I/O error, dev sdc, sector 181463392 op 0x1:(WRITE) flags 0x0 phys_seg 4 prio class 0

 md: disk3 write error, sector=181463328

 

I rebooted, took the disk out of the array, added it back in and did a data re-build. Less than 24 hours and it has happened again to the same disk. SMART reports no issues. I've attached diagnostics. Any idea what this could be?

 

Any help would be very appreciated.

 

fatboy-diagnostics-20220812-1900.zip

[JorgeB]

It's not logged as a disk problem, suggest you replace both power and SATA cables to rule that out (or slot) and try again.

[paulmorabi]

On 8/12/2022 at 7:05 PM, JorgeB said:

It's not logged as a disk problem, suggest you replace both power and SATA cables to rule that out (or slot) and try again.

 

Hi,

 

After a few days of testing, I've tried the following:

* The disk is connected to an LSI card. I have 2x 4-way data cables for connecting to disks. I swapped the data cable for a spare on the other cable. Still got the same error.

* I have a power cable splitter that has 4x cables for connecting to disks. I swapped the power cable for an unused cable. Same error.

* Replaced the power splitter cable entirely. Same error.

* Plugged the disk into a free SATA power cable plug coming from the PSU. Same error.

 

The error is:

 

Quote

blk_update_request: I/O error, dev sdc, sector 1221470944 op 0x1:(WRITE) flags 0x0 phys_seg 1 prio class 0
md: disk3 write error, sector=1221470880

 

 

When it first occurred on Thursday, the next failure was nearly 24 hours later. It's now occurring at increasingly smaller intervals, down to only 3-4 hours.

 

SMART extended self test reports no issues.

 

Nothing else has changed in terms of hardware since the end of last year and between then and last week, this was never an issue.

 

Could it be the disk or should I investigate further?

 

[JorgeB]

I would just replace the disk, SMART cannot detect all issues.