So I've been having an issue since installing a new SAS HBA (9305-24i in IT mode) one of the ports is flakey and two disks connected to that port will have UDMA CRC errors tick up (1st and 4th disks on the breakout). I confirmed it wasn't the cable by replacing with a new cable. I confirmed its not the disks by swapping that entire branch with another. The issue follows the port on the HBA. This has left me in a "precarious" situation where one of my parity disks is disabled and one data disk is disabled. The issue was initially resolved after swapping the ports on the HBA so I chocked it up to "Maybe it was just a loose seat" and now I'm nowhere near the server to do any maintenance, and there's nobody else who can do it for me.

I plan on powering off the server until a new HBA arrives to replace this one - but until I am able to migrate some services around I can't really do that. Since I have dual-parity, the array is still emulated, but if I lose any one disk I'm kind of SOL on recovery, correct? So far this is the only indicated hardware issue I have.

All unaffected disks should still be OK since they are independent. The affected disks, if not actually bad, would be out-of-sync and would have no writes after they became disabled since those writes are only emulated by the rest of the in-sync array. The less you write to the disabled disks the less out-of-sync they would be, and those writes would be lost anyway if they couldn't be emulated any longer. Also lost writes could cause corruption (missed filesystem metadata updates) and require filesystem repair.