Crash - "Data Storage Interrupt, Processor"
Christopher Davey
Model 320H AIX 3.2.0
Has anyone had a crash like this before ? The only significant
change to the machine in the last six months is that a 2GB Seagate
disk was added 5 days ago.
Help
Chris
--
----WNA----------------------------------------------------------------------
Chris Davey Internet: ch...@quay.ie
Quay Financial Software Phone : +353 1 612377 Fax: +353 1 607592
Christophe Wolfhugel
I have this problem whereas over machines in the campus neighborhood don't.
The cause has never been clearly isolated. But it is definitely related
to a bad interaction with power-supply events. And SCSI is maybe to
sensible.
When we did not have the UPS IBM told it was out power supply. Errors
were only on the internal drive or on the (IBM) streamer.
Ok, we installed a UPS (1.4Kva) which according to our IBM SE is enuff.
I'm not convinced on this. As the problems started again and again
we wanted to check that point. A local operator plugged the UPS out for a
few seconds. Errlog showed SCSI errors on the internal disk, the printer
ejected a sheet of paper. Still nothing with the external disk.
So either the UPS is in trouble or the 320's power supplies are too weak
and too sensible to small variations. No more precise idea yet.
The PC-RT which was on the same place never crashed due to electrical
problems, even without the UPS.
--
Christophe Wolfhugel | Email: Christophe...@grasp.insa-lyon.fr
Matt Accapadi
>
>Model 320H AIX 3.2.0
>
>Has anyone had a crash like this before ? The only significant
>change to the machine in the last six months is that a 2GB Seagate
>disk was added 5 days ago.
>
>Help
>Chris
This can be caused by any one of a million things. You need to look at
the dump to determine the cause. First execute "sysdumpdev -L".
That will tell you the dump device it dumped to and whether it was successful
at dumping it. Then dd the dump image off the dump device - it's a lot faster
than running crash on the dump device directly. For example,
dd if=/dev/hd7 of=dumpfile
where in this case /dev/hd7 was the dump device. The output file (in this
case I called it dumpfile) will be the same size as the size of your dump
logical volume (ie. /dev/hd7), so make sure you have enough room for this.
Then run "crash dumpfile". Once within crash, there are many things you could
do depending on the crash. You might start by typing in "trace" or "t". That
will give you a stack trace for the currently running process. Type in
"user" or "u" for more detailed information on the current process.
You'll have to call support or post what you've found so far to determine
what to do from here on.
--
Matt Accapadi Internet: acca...@austin.ibm.com
IBM AWD AIX System Performance VNET: ACCAPADI at AUSTIN
11400 Burnet Rd. Bld. 906/9632 yaknet: 512-838-3193 Tie 678-3193
Austin, TX 78758-3493 fax: 512-838-8344
John F Haugh II
>The cause has never been clearly isolated. But it is definitely related
>to a bad interaction with power-supply events. And SCSI is maybe to
>sensible.
What is the physical relationship of the SCSI cable to the power cord
or other power supply lines? The reason I ask about the external SCSI
cable is the possibility of some noise being induced on the external
cable affecting the internal devices.
On another note, if you are convinced it is the physical location of
the system, move the system. Swap a working machine in another location
for the mystery machine. What happens so far as the error moving will
tell you where the fault lies.
--
John F. Haugh II | Quality is ... knowing who | MaBellNet: (512) 823-8817
SneakerNet: 042/2F068 | your customer is and what | VNET: HAUGH at AUSVM8
[ DoF #17 ] [ TSAKC ] | your customer wants. | Disc: I speak 4 me, !IBM.
Sebastian Kremer
The machine wouldn't boot and stop with a LED code that according to
the manual meant `Data Storage Interrupt'.
The PC and RS/6000 keyboards looked virtually identical, for the
RS/6000 the cable is quite a bit longer.
--
---------------------------------------------------------------------------
Sebastian Kremer <s...@thp.Uni-Koeln.DE>
Institute of Theoretical Physics, University of Cologne, F. R. of Germany
My new job is without email access (yet). I can read my mail only
occasionally. The mail address will continue to work, but don't
expect answers in less than a week.
---------------------------------------------------------------------------