Vendor: QUANTUM Model: FIREBALL ST4.3S Rev: 0F0J
I got this firmware update from Quantum tech support about a month ago. It did
not help.
Peter Svensson wrote:
>
> The most current firmware revision also have a problem with a tagged-queue
> depth of more than 7 under heavy load.
Though it will die from load, it normally goes into EATA resets during pretty
light load as if it has a 'wakeup' problem.
I'm planning to replacing the drive soon. The thing worring me is the opps.
I'm seeing the opps more then the resets right now.
> > occasionaly see EATA-Resets. I was replaced the mobo and that
> > seemed to produce Opps and a higher degree of EATA-RESETS. The 3
> > other boxes I have with this same board run like a rock (but run
> > OS/2 and DRDOS, not linux)
>
> If the changed board is the only one that goes Oops Id guess you swapped
> a good for bad board ?
Just before the board change happened the system was put into full service. I
*might* have seen an opps with the old board. Before that the system was running
for several months with only the occational eata reset...but of course this was
with practically no load.
FYI the old board was an ABIT IT5H. Intel HX chipset. This one had a Cyrix
P200MX.
Also FYI, a system sitting right next to it is a dual P100, with a 3.2GB quantum
fireball. Tyan Titan 3 96MB. Hell of a time getting memory above 64MB working
but once I did it runs very good. I DO see EATA resets on it, but they are so
rare it's not a major concern. Right now this box has been up 9 days and has
assumed the load that should be going to the other box.
I have swapped the DPT's around. No change.
> Is that board one of those that overclocks the PCI bus to 37.5MHz when you
> run 75MHz bus ?
Yes it runs the PCI sync/2. (Which is the point to the faster bus IMO)
The PCI bus contains 2 SVEC PN100TX and right now a DPT 2144W.
This DPT was previously in my personal system up until a few months ago. That
has a custom 90MHz bus (45MHz PCI) and the HBA ran flawlessly for over a year.
The Aladdin chipset is rated to 83MHz. I doubt any of this is brought on by bus
speed.
One thing I left out of the opps report....I'll see stuff like this:
free_one_pmd: bad directory entry 01000000
in dmesg sometimes. Once that's present the box has a few hours left in it at
best before an opps.
-- http://www.psychosis.com/emc/ Elite MicroComputers 732-541-4214 http://www.psychosis.com/linux-router/ Linux Router Project "The only way to remain cutting edge, is to design for things that do not yet exist." -- Me : )- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu