Thanks to the people who responded so fast:
- Dr. Tom Blinn, 603-884-0646 [tpb_at_doctor.zk3.dec.com]
- Jim Fitzmaurice [jpfitz_at_fnal.gov]
- alan_at_nabeth.cxo.dec.com <mailto:alan_at_nabeth.cxo.dec.com>
My original question will be in the end, so will be the answers I got from
the people mentioned above.
After reading the replies and some testing I found out that when I connect
the SCSI cable to the second controller I can see my disks again. Because of
the address change I'm not able to boot from the disks now. Anyhow now I'm
convinced that my data is still on the disks and it's not some disk crash, I
will call Service on monday to have the controller replaced.
Alan wrote:
I think the KZPCA is low-voltage differential, while the
KZPBA-CX is single-ended. I read that correctly functioning LVD can pretend
to be single-ended, but I've never tried it. The last person I heard of
trying it (using an IBM LVD disk) didn't have much luck. I don't know what
adapter or system he was using but he ended up getting a LVD adapter for the
disk.
There may also be significant differences in the disk
packaging that will present moving disks from one bus to the other; first
generation StorageWorks vs. 2nd generation or some variant of Compaq
packaging.
While the problem with the B bus could be a bad adapter, it
could also be a cable problem, termination, bad shelf, insane device, etc.
Check for any obvious problems and then remove all the devices from the bus,
repeating the system init as each device is added. If it works and then
doesn't with the addition of a device, suspect that device and continue with
the others. Then try the suspect device by itself. If it works add others
until it fails.
Jim wrote:
Marco,
My advice is to stop doing daily reboots! I've had uptimes
of 180 days or more and have only done reboots due to, power outages,
hardware failures, or to change kernel parameters. My systems rarely
experience any problems of any kind.
Dr. Tom Blinn wrote:
Although it MAY be a failing SCSI controller, in my
experience, the "waiting for adapter to poll" is often due to bad cabling or
a bad terminator.
If you have nothing connected to the other controller, you
might be able to just move the drives. Without knowing more details about
your configuration than I really want to know, I can't easily tell you
whether that will work. And it may make the system unbootable if you're
running a version in the V4.0x stream, since you might change the disk IDs
and hence their names, and either the device files will not exist or they
won't match up to your file system names in /etc/fstab..
Original message:
The DS10 of one of our customers suddenly don't want to boot anymore. I
suspect (hope) it's the SCSI controller which is failing.
When booting the screen display:
os_type: UNIX - console CIPCA driver not started
Testing the System
waiting for pkb0.7.0.17.0 to poll
...
...
...
waiting for pkb0.7.0.17.0 to poll
Configuration:
PCI slot 2 KZPCA-AX
PCI slot 4 KZPBA-CX
sho dev:
dva0.0.0.0.0 DVA0
pka0.7.0.15.0 PKA0 SCSI Bus ID 7
pkb0.7.0.17.0 PKB0 SCSI Bus ID 7
There are 2 disks connected to the KZPBA-CX, sho dev will not show them and
the system is not booting. Already checked all cables and the system was
running before the daily reboot.
My conclusion is that the SCSI-controler is not functioning correctly
anymore. Can I connect my drives to the second controler? Anybody any
suggestions?
TIA
With regards,
Marco van Herwaarden
Engineering Manager Woningbeheer
Inter Access Cens B.V.
Groningenweg 6
2803 PV Gouda
the Netherlands
Tel. +31 (0182) 537477
Mail: marcoh64tru64_at_yahoo.com <mailto:marcoh64tru64_at_yahoo.com>
Received on Fri Jun 30 2000 - 18:41:54 NZST