Re: [gentoo-user] struggles with SATA

["Kent Fredric" <kentfredric@gmail.com>]

On 1/11/07, Alan E. Davis <lngndvs@gmail.com> wrote:
> Using an AsRock 939SLI32-eSATA2 motherboard and a Western Digital sata
> drive, I have been having alot of trouble.  I found AHCI driver to work, at
> least detect the drive/partitions.  I have had no end of problems.
>
> I had moved this drive from another machine w/ a Tyan motherboard, where it
> worked flawlessly.  Most recently, it is impossible to write to the drive
> (at least alot of times): it causes what seems like a lockup.  I cannot find
> anything about this.  I compiled AHCI support into gentoo-sources-2.6.19-r2.
>
>
> I want to ask if some kind soul can point me in a useful direction?

I've suffered a lot of hell at the hands of SATA. I was experiecing
what you describe, drive lockups, IO write drops, and even drive
power-up-power-down cycling. But, i think after several months I think
I've found it.
My troubles started becoming more clear what they were when i turned
on full APIC support in the kernel, which I had disabled at an earlier
date due to weird "nvidia+xcomposite=system failure" issue, which
appears to have been resolved.
After i turned on APIC, i noticed dmesg started reporting drive
add/deletes vaugely reminiscient of USB key  plugging. So It dawned on
me, that maybe, I finally had full SATA-HOTSWAP support working
properly.
So, I wondered why my drives were tripping hotswap events, and after a
bit of mucking around, I concluded, that it was a loose cabling issue
in the case somewhere.

So I played around with touching various cables, and noticed with some
of them, all i needed to to was -touch- the cable for the drive to do
a nasty "power up/powerdown" jump, and then replaced some of my sata (
i had spares made by a different manufacturer ) cables, some of my
sata-2-molex power adaptors, and have since found my problems have
subsided :)

So my advice, is give the cables a bit of a jiggle while the computer
is running and see if you can cause a failure event to occur by doing
so. :)

( As a side note, I now distrust SEAGATE SATA drives, they seem to run
10°C hotter than all my other drives ( Hitachis ) and seem to easily
hit 60°C at which point they also start failing and having the
SMART_ERROR_RATE being so huge it keeps going into integer overflow. )

Best of luck, I know how nasty it can be having drives jumping foot
all the time. I have 4 drives. All SATA :P.

-
Kent

-- 
gentoo-user@gentoo.org mailing list