Disk errors

[NickT]

I just had an issue with a seagate exos 14tb parity drive causing the following errors and system lockups:

 

Quote

Aug 29 05:03:52 Paragon-Unraid kernel: ata6.00: failed command: READ FPDMA QUEUED
Aug 29 05:03:52 Paragon-Unraid kernel: ata6.00: cmd 60/40:80:60:b4:08/00:00:28:00:00/40 tag 16 ncq dma 32768 in
Aug 29 05:03:52 Paragon-Unraid kernel:         res 40/00:90:e8:b3:08/00:00:28:00:00/40 Emask 0x50 (ATA bus error)

 

After swapping cables and ports, it didn't go away. So I removed the drive and all seemed well. Now another identical drive is showing read errors during a read check. Both drives are only a few months old. I ordered a replacement for the parity drive, of the same model, and now during pre-clear the syslog is showing errors on that too. It seems pretty unlikely that I'd have 3 bad drives in a row.

 

My other drives, connected to the same HBA seem to be fine. Could this be a compatibiliy issue?

 

 

paragon-unraid-diagnostics-20220901-1513.zip

[JorgeB]

Try connecting those the the onboard SATA controller to see if there's any difference.

[NickT]

Just did that, with new cables. So far the preclear is going much faster with no errors. First time it slowed as far as 29 MB/s, now its going around 280. I've started a read check on the array, I'll update with the result of that.

 

The HBA works fine with the other drives,  though they are on a different cable. 

[JorgeB]

Could be some firmware compatibility issue, with the drives or the LSI.

[NickT]

Started getting errors on the new drive that's preclearing.

paragon-unraid-diagnostics-20220902-1318.zip

[NickT]

The power supply in this system is quite old. Could that be the issue?

[JorgeB]

Just now, NickT said:

Could that be the issue?

It could.

[NickT]

Ok, I'll swap that out and try again.

[NickT]

The new power supply seems to have solved it.