SOLVED Disks on PCIe-SATA board failing on Scrub

[cK7Tcosta]

Hello all,

I think I need some help understanding the issues I've been having with my disks.
I have a Dell T30 with 32Gb RAM and Xeon E3-1225 V5 and an ORICO 5xSATA PCI card (this might be relevant)

In my old setup I was running TrueNAS Core on a mechanical HD on SATA0, then had a mirror with two 4TB IronWolf PRO on SATA1 and SATA2 (pool name = big_storage) and finally a RAIDZ1 with 4x Samsung 320Gb HDD (pool name= small_storage) on SATA3 + 3 of the ORICO ports. big_storage was for backups via SMB and small_storage had jails and nextCloud.
This setup ran perfectly for months but, after experimenting with Scale on another machine, I wanted to move to Scale. I also got some extra bits of hardware and wanted to improve the system.
So now I'm running Scale 22.12.3.2 and the hard drive config is:
SATA0 - SSD 128 GB - boot_pool
SATA1 - SSD 128 GB - jails_pool (for apps and VMs)
SATA2 - HDD 4TB - /dev/sdc
SATA3 - HDD 4TB - /dev/sdd
Orico0 - HDD 2TB - /dev/sde
Orico1 - HDD 320GB - /dev/sdf
Orico2 - HDD 320GB - /dev/sdg
Orico3 - HDD 320GB - /dev/sdh
the HDDs are in a strippe of 3 mirrors called big_pool and the 2 TB disk is waiting for me to get another 2 TB to have a symmetrical mirror.

Ok, all that being said, after a few days I started having the errors bellow on disks e to h. At first I thought it was related to having Scrub on Sundays at midnight and having SMART tests starting at 2am and thus having them overlap. However I deleted the scheduled test and that did not solve.

Yesterday I ran SMART test on all drives and they all passed, plus I did a Scrub on the pool and everything worked. Before going to sleep I deleted my backup data-sets (that were damaged) and reconfigured everything. As you can see from the times above, this morning the poop hit the fan.
As for before, if I just reboot the server everything is OK, no errors, everything online.

Could it be that the Orico PCI card does not have enough horse power to be used in a pool that also has disks on the mainboard?
Could it be the mismatched disks?
The thing is, the disks were also mismatched and I also had the ORICO before, could ir be that I had two pools and now they are all on one pool?

I'm really lost here and would appreciate any help

 

[unseen]

ORICO 5xSATA PCI card (this might be relevant)

It is indeed relevant. This controller is based on the JMB585 chip from JMicron and is known not to work properly with TrueNAS.

 

[cK7Tcosta]

It is indeed relevant. This controller is based on the JMB585 chip from JMicron and is known not to work properly with TrueNAS.

Ah... well thanks for the information.
I'm going to look a bit more into this because it's looking like it might be a Scale issue since in Core I never had any.
It sucks a bit because I wanted to use docker but if I need to go back to Core I'll do it

 

[cK7Tcosta]

So, as a head's up for future readers, I replaced the SATA - PCIe board for a new one that was already on the mail and, at least for now, it seems the problem is solved.
The one I have now is Orico PSA-M4U and it's based on an Intel Chipset so it seems the chipset was really the issue.

Thanks for the help