HDD Failing... But There's Nothing Wrong With It.

**Bushflyr** · June 11th, 2012

Thanks, I'll fire off a nastygram to Jetway and see if they can help me test it somehow.

**Bushflyr** · June 12th, 2012

Possible problem found!!! (hopefully) I took the whole machine apart and started futzing around. I had the SATA cables beautifully routed and bundled but when I started wiggling them they looked sort of loose in their sockets on the MOBO. Not because they were loose per se, but because they don't fit that well. I cut all the ties and ran them loose to try and alleviate any stress on the sockets. Rebooted and /dev/md127 came up clean.

Stopped md127 and did a quick:

Code:

sudo mdadm --create --verbose --assume-clean /dev/md0 --level=raid6 --raid-devices=4 /dev/sdb1 /dev/sdc1 /dev/sdd1 /dev/sde1

and it seemed to work.

Back to stress test again. And order some good locking SATA cables.

I'm hopeful.

**rubylaser** · June 12th, 2012

Geez, cable management is a great thing... except when it makes your connections inconsistent

I really hope this solves your problem.

**Bushflyr** · June 12th, 2012

Well. Crap.

I can drop a single instance of my music directory on it and it seems to do OK, but when I drop it into 2 separate directories at the same time the RAID starts failing disks. Back to the drawing board. New cables should be here in a couple days in any case.

**rubylaser** · June 12th, 2012

Can you setup a partition and filesystem on each disk and mount them individually and copy to them (or better, try out timemachine on each) and see if you see the disks become unavailable?

**linkageless** · June 12th, 2012

Also, try a RAID-1 on just a couple of those disks and the other two removed. I know it's no use capacity-wise for you, but at least you'll have an idea what's reliable and what isn't.

If two disks (plus system drive) work ok, then try adding another and test again.

**Bushflyr** · June 12th, 2012

Both excellent suggestions. Thanks. Considering how many times I've rebuilt this thing already doing individual disks will be simple. After that 2x raid 1 then add in more disks and see what happens. And migrate to AHCI while I'm at it.

The only saving grace through this whole thing is that the OS is on a separate drive that's been stable. I would have shot this box a month ago if I had to redo the OS every time.