EQ12 SATA CRC errors with 2TB Samsung 870 EVO

pgdh

I have 2x EQ12 16G/500G bought from Amazon, both running Proxmox.

One has a 1TB Samsung 870 EVO SATA drive with no issues.
The other has a 2TB Samsung 870 EVO SATA drive which sees many errors like this ...

[Tue Sep 19 09:58:02 2023] ata2.00: exception Emask 0x10 SAct 0x18000 SErr 0x400100 action 0x6 frozen
[Tue Sep 19 09:58:02 2023] ata2.00: irq_stat 0x08000000, interface fatal error
[Tue Sep 19 09:58:02 2023] ata2: SError: { UnrecovData Handshk }
[Tue Sep 19 09:58:02 2023] ata2.00: failed command: WRITE FPDMA QUEUED
[Tue Sep 19 09:58:02 2023] ata2.00: cmd 61/f0:78:f8:06:34/00:00:38:00:00/40 tag 15 ncq dma 122880 out
                                    res 40/00:01:00:4f:c2/00:00:00:00:00/00 Emask 0x10 (ATA bus error)
[Tue Sep 19 09:58:02 2023] ata2.00: status: { DRDY }
[Tue Sep 19 09:58:02 2023] ata2.00: failed command: WRITE FPDMA QUEUED
[Tue Sep 19 09:58:02 2023] ata2.00: cmd 61/a8:80:e8:07:34/00:00:38:00:00/40 tag 16 ncq dma 86016 out
                                    res 40/00:01:06:4f:c2/00:00:00:00:00/00 Emask 0x10 (ATA bus error)
[Tue Sep 19 09:58:02 2023] ata2.00: status: { DRDY }
[Tue Sep 19 09:58:02 2023] ata2: hard resetting link
[Tue Sep 19 09:58:02 2023] ata2: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[Tue Sep 19 09:58:02 2023] ata2.00: supports DRM functions and may not be fully accessible
[Tue Sep 19 09:58:02 2023] ata2.00: supports DRM functions and may not be fully accessible
[Tue Sep 19 09:58:02 2023] ata2.00: configured for UDMA/133
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#15 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#15 Sense Key : Illegal Request [current]
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#15 Add. Sense: Unaligned write command
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#15 CDB: Write(10) 2a 00 38 34 06 f8 00 00 f0 00
[Tue Sep 19 09:58:02 2023] I/O error, dev sda, sector 942933752 op 0x1:(WRITE) flags 0x700 phys_seg 7 prio class 2
[Tue Sep 19 09:58:02 2023] zio pool=rpool vdev=/dev/disk/by-id/ata-Samsung_SSD_870_EVO_2TB_S754NX0W713335N-part3 error=5 type=2 offset=481707290624 size=122880 flags=40080c80
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#16 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#16 Sense Key : Illegal Request [current]
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#16 Add. Sense: Unaligned write command
[Tue Sep 19 09:58:02 2023] sd 1:0:0:0: [sda] tag#16 CDB: Write(10) 2a 00 38 34 07 e8 00 00 a8 00
[Tue Sep 19 09:58:02 2023] I/O error, dev sda, sector 942933992 op 0x1:(WRITE) flags 0x700 phys_seg 13 prio class 2
[Tue Sep 19 09:58:02 2023] zio pool=rpool vdev=/dev/disk/by-id/ata-Samsung_SSD_870_EVO_2TB_S754NX0W713335N-part3 error=5 type=2 offset=481707413504 size=86016 flags=40080c80
[Tue Sep 19 09:58:02 2023] ata2: EH complete
[Tue Sep 19 09:58:02 2023] ata2.00: Enabling discard_zeroes_data

It appears to be a known issue with the Bee-Link's use of unshielded SATA ribbon cables.

I am going to try a Crucial MX500 (as in my experience, they are more immune to noise).But I think the issue really is down to Bee-Link's design.

It appears to be an issue only at 6Gbps.
When I boot the kernel with 'libata.force=3.0Gbps', the problem seems to go away (but the performance is poor).

Do you have a better cable option?

I have wondered about applying some adhesive metallic tape to the cable.
Is this likely to help?

pgdh

UPDATE...

The 2TB Crucial MX500 has been running fine. I have written over 600GB to it without a single CRC error.
The 2TB Samsung 807 EVO worked ok when restricted to 3Gbps, and works ok using a SATA to USB adaptor.

Two theories ...

1) the Samsung is more sensitive to noise on the SATA bus
2) the Samsung generates more noise which gets picked up by the SATA bus

Either way, I can't help thinking that the issue is with the unshielded SATA ribbon cable in the EQ12.