[Pc_Support] 3ware corruption on K7S5A

[Jason Boxman]

On Wednesday 15 February 2006 12:01, Bryan J. Smith wrote:
> Jason Boxman <jasonb at edseek.com> wrote:
> > Months later, I thought I'd Google today for the heck of it and
> > found this where someone claims confirmation from 3Ware back in
> > 2002 that this is a problem with the K7S5A.
> > http://www.geek.com/news/geeknews/2001nov/bch20011129009082.htm
>
> I stopped using advanced storage and other multi-function/bridged
> devices on the SiS735 chipset after I found APIC/I2C comaptibility
> issues.  E.g., I had issues with PCI-to-CardBus controllers.
>
> But I still liked the K7S5A as a general desktop.  I even ran it with
> a 3Ware Escalade 6410 card for awhile.  Kinda wondering now if that's
> why I lost that old XFS 1.0 version /var filesystem years ago -- the
> combination of a 3Ware not flushing a buffer and the XFS bug's
> inability to recover such an inconsistent filesystem (fixed in XFS 1.1).

Heh.  Turns out the 3Ware board in question seems to be dead.  However, it 
fails differently than it did in the K7S5A.  What's more, it failed 
identically to another 6200 board that is known good in a different machine 
for over a year.  So it seems this 6200 was probably good in my K7S5A, but 
unusable thanks to ECS.

Now, it's just plain dead.

Sigh.

Now I'm accessing my ATA disk over the ServerWorks IIILE directly and it's so 
slow at the mdma 2 max. ;)

/dev/hda:
 Timing cached reads:   812 MB in  2.00 seconds = 406.00 MB/sec
 Timing buffered disk reads:   42 MB in  3.07 seconds =  13.68 MB/sec

-- 

Jason Boxman
http://edseek.com/ - Linux and FOSS stuff