Troubleshooting Pool Errors

[ere109]

A few months ago, my onboard LSI 2308 HBA started throwing massive amounts of errors on a five-drive ZFS raidz2 pool, specifically under heavy write load. I asked for community help, and after updating firmware and changing cables, I was told to swap to the onboard sata controller. The problem seemed to go away (during this time I updated from 11.1 to -U1 to -U2 to -U3).
I sent the motherboard into Supermicro for repair.
The box just came back yesterday and I got really excited. Then I opened it. They tested and reported no problem in tests on RHEL 7.3 and Windows 10. It looks like they only verified that drives were populated, but may not have put stress on the controller.
Now I'm in a pickle. I don't want to reassemble everything only to have the same issue come up again.
My current thoughts:
My controller still had IR firmware. Raid is disabled in bios, but I've read of people still having problems until they install IT.
The drives are brand new Seagate 8TB Ironwolf, but I'm trying to identify all possible factors.
I've also read a few threads about instability of the FN SAS driver with various firmware combinations.
I could send it back for harder testing, but suspect I'll get the same result (whether they actually perform further testing or not).
I'd love some new thoughts and feedback. I feel a bit disappointed today.

 

[Ericloewe]

Are you sure it wasn't simply overheating? They do need a bit of airflow over the heatsink.

 

[ere109]

Hmmm... I've got two 120mm fans blowing in from the front and right across the board. I don't have anything directly aimed at the controller heatsink. I have recently purchased an active 4U CPU cooler, and that may dissipate heat that's currently spreading out from the processor to other components...