[rescue] Vorpal Dustbunnies from Heck!
bear at typewritten.org
Fri Aug 12 21:04:34 CDT 2005
On Aug 12, 2005, at 5:00 PM, Phil Stracchino wrote:
> It's all cleaned out now, and re-initializing RAID5 on ten disks. I
> figure on running it for a while with a two-disk hot spare pool and
> how it goes. If that goes well, I may add an extra disk into it
> and cut
> the hot spare pool down to one.
Both your 711 enclosures are UltraSCSI rated (6-slot and say
"UltraSCSI" in the front lower-left corner), right?
If not, and cleaning out the dust and getting things cooled down in
the enclosure doesn't get rid of your problem, you're probably
running into a cabling issue. The non-UltraSCSI enclosures do not
reliably run at 20 MHz, regardless of whatever anecdotal evidence
somebody else may provide; the wiring in the backplane exceeds the
maximum allowed for 20 MHz operation. Enclosures marked "UltraSCSI"
don't have this problem.
The SunSwift PCI has an UltraSCSI interface, and the DRHS drives are
all UltraSCSI capable, so the target-initiated synchronous transfer
rate negotiation causes the bus to run faster than the backplane can
handle. There's a driver parameter you can adjust to prevent the bus
from operating faster than 10 MHz, but I can't recall it off the top
of my head.
If you have an UltraSCSI 711, then it's not the problem... unless
you've got a mile of cable between your HA and the 711 (@:
More information about the rescue