Multiple errors on < 1yo SeaGate 8TB HDD since 6.9 upgrade

[SomeRandomSod]

Hello all,

Today I saw on my server in the /Main tab 896 errors on one of my hard drives. I narrowed down these errors to 01h30 on the 03/03/2021.
1h30 of the first Wednesday of the month is when I run my monthly parity check, so it was during the start of the parity check that the errors popped up.

The errors concern disk 1. In the logs this repeats multiple times :

 

Quote

Mar 3 01:30:21 Tower kernel: sd 3:0:5:0: [sdi] tag#89 Sense Key : 0x2 [current]
Mar 3 01:30:21 Tower kernel: sd 3:0:5:0: [sdi] tag#89 ASC=0x4 ASCQ=0x0
Mar 3 01:30:21 Tower kernel: sd 3:0:5:0: [sdi] tag#89 CDB: opcode=0x88 88 00 00 00 00 00 00 00 10 40 00 00 04 00 00 00
Mar 3 01:30:21 Tower kernel: blk_update_request: I/O error, dev sdi, sector 4160 op 0x0:(READ) flags 0x4000 phys_seg 128 prio class 0

The device is relatively new, less than a year old. It's a 8TB Seagate Ironwolf drive. I set an extended SMART test to run, seems to be stuck at 10% for the last 2h or so.

I attached the logs to this post.
 

Any help is appreciated,
Regards,
George.

tower-diagnostics-20210304-1325.zip

Edited March 4, 2021 by SomeRandomSod

[JorgeB]

There are knwon issues with LSI + 8TB Ironwolf (possibly other capacities also), one workaround is to use the onboard SATA ports for those, until there's a new LSI driver.

[SomeRandomSod]

Ty for the answer, is this to do with the 6.9 update? If so I'll probably just roll back, my unraid server is a good 10h drive away and to switch the ports would be a pain.

[JorgeB]

1 hour ago, SomeRandomSod said:

is this to do with the 6.9 update?

Yes, it started with the v6.9-rcs, likely due to a newer LSI driver.

[SomeRandomSod]

Thanks. Quick question, is this more of a "technical" error or is there an actual chance of data loss?

[JorgeB]

It's disabling disks for some users, so it can cause data loss.

[SomeRandomSod]

Can confirm, it disabled one of my disks today.
Reverted back to 6.8.3 and rebuilding parity now.

Hopefully this will be fixed pretty soon, I really needed those multiple pools

 

[limetech]

On 3/4/2021 at 8:12 AM, JorgeB said:

Yes, it started with the v6.9-rcs, likely due to a newer LSI driver.

 

Hi @JorgeB - do you know of a kernel bug report related to this?

[JorgeB]

11 hours ago, limetech said:

do you know of a kernel bug report related to this?

Nope, just that there were a handful of cases reported, IIRC always with the ST8000VN004 model, though other capacities might have the same problem, and always fixed by downgrading to v6.8 or moving those disks to a different controller.

[Squid]

On 3/5/2021 at 2:56 PM, limetech said:

 

Hi @JorgeB - do you know of a kernel bug report related to this?

 

@limetech