I'm experiencing multiple errors, and Unraid doesn't seem able to unmount the drives. The syslog is full of exceptions and errors. I've been experiencing drive errors on and off for a few months now, and I recently replaced all of the power and sata cables with lockable ones. When the problems persisted, I thought there problem might have been a fault with the two Promise TX4 sata cards, but this time the errors seem to be across all drives, including those connected directly to the motherboard. Any thoughts would be appreciated. 

 

Syslog is here: http://pastebin.com/m5a626f87

This may or may not be a coincidence, but the 3 drives that are most involved in the errors are your 3 Seagate ST3750640AS 750GB drives - Disk 1 (sdc, second Promise card, firmware 3.AAE), Disk 2 (sdh, motherboard port, firmware 3.AAD), Disk 7 (sdi, motherboard port, firmware 3.AAE).  They are also the 3 drives with Seagate SATA150 jumpers still installed, see the Improving unRAID Performance, Remove SATA150 Jumper section.

 

I cannot say whether there is a problem with the SATA limiting jumper, because I have never seen or heard of an issue like this with them.  There may be a firmware issue with these 3 drives, I don't know, but might be worth checking out.

 

Otherwise, I would suspect your power supply.  Do you have another to substitute as a test?  The fact that so many drives are involved with no obvious common factor makes the power supply a suspect.

 

A very odd fact that I don't understand is that the 3 drives, and sometimes additional drives too, would have problems *immediately* AFTER spinning down, which makes no sense at all to me so far.  I'm out of time for any further analysis at the moment ...

Thanks Rob, I'll dig out another psu and see if that helps. Currently it has a Thermaltake Toughpower 750w (http://www.thermaltake.com/PRODUCT/Power/ToughPower/W0116/w0116.asp). I am running five drives off each of the two sata rails, which are rated at 18A each. So, on the face of it the psu should have ample capacity, unless it's faulty.