This is the second month in a row that the scheduled parity check failed and Disk 5 was marked with a Red-X and removed from the array. I have been monitoring the system log all month and there have been no errors and the array was performing fine.

 

When this happens, the drive itself exhibits no hardware issue. I have used 2 different drives in this slot, and if the drives are wiped and rewritten, they work fine. So it does not appear to be a drive error. 

 

Parity Check starts at 1:00am. The log shows the first error happening at 04:21:45 (ata 9 is a Crucial 256GB SSD in the cache slot)

 

May 1 04:21:45 Tower kernel: ata9.00: exception Emask 0x0 SAct 0x2000 SErr 0x0 action 0x6 frozen

 

I am attaching the system log and the diagnostics output (I have not rebooted or done anything except read the logs)

 

If anyone can shed some light on this, I would be grateful.

 

Thanks.

 

tower-syslog-20170501-1028.zip

tower-diagnostics-20170501-1007.zip

Problem is the Marvell controller, try disabling vt-d if not needed, check for a board bios update and/or use the controller in a different slot if available, if all fails get a non Marvell controller.

Thank you for your response. Can you recommend a 4 or 8 port controller card?

Get one with a LSI chipset, e.g., 9211 or 9210-8i, most by the cheaper clones, like the Dell h310 or IBM m1015 but these need to be crossflashed to work with unRAID.

Thanks. I ordered an LSI 9211-8i flashed to IT mode.  

 

One last question. Should I wait until the new card is installed to replace the drive? In the past, I have been able to rebuild the failed drive and it seems to work fine until the next parity check. As mentioned, I have been monitoring the system log since last month, and no indication of any errors.

 

 

Rebuild and wait for the new controller before running another check.

Solved. I received the LSI 9211-8i, flashed it to IT mode, installed and ran parity check to completion with no errors.

 

Thanks johnnie.black for the help.