HP DAT72 and disks on same bus on a PWS

[Rob Leadbeater]

Hi,

Which Qlogic card are you using specifically ?

> but once booted the tape was missing.

By missing do you mean the /dev/tz* device wasn't present ? (I'm assuming you're using 4.0x). Is it the case that you just need to create the device files with MAKEDEV ?

What does the output of "scu show edt" show you ?

Cheers,

Rob

[Bengt Nilsson_2]

Well, I was changing my setup without telling, sorry.

This is the situation:

I have now confirmed that the tape and at least two disks can work together on the internal scsi bus provided that "Force Single-Ended mode" is enabled on all disks, using tru64 V4.0F.

I have now

dkc0
dkc100
mkc500

and nothing connected to my LSI Logic KZPCA-AX interface for the moment. My QLogic KZPBR-CY did not work, so I put it aside.

After booting a pre-installed V5.1B-4 disk (installed without the tape or tape connected to a different scsi configuration) I am not able to get the tape into the system. genvmunix/doconfig is NOT doing the trick. Someone suggested it is necessary to REINSTALL V5.1B to detect and get new hardware support. Is this really true?
./MAKEDEV tz21 did not produce any working nrmt0h device. # file /dev/nrmt0h claims scsi id#0 which is not true.

Should some other tool be used in V5.1B to install new hardware?

For now, my prime target is to vdump my 5.1B-4 disk on tape, since I put a lot of work in the configuration of some external hardware support, and I need this physical disk to replace a failing 4.0F system disk on another "production" machine.

I tried to do the vdump operation from the 4.0F disk, but strangely enough, I can mount the /usr/users file system on the V5.1B-4 disk but not / or /usr, Mount claims they do not contain a valid AdvFS file system, while disklabel -l says it does.
So I guess I have to boot on the 5.1B-4 disk to do the vdump, but now I have no tape support in it.

[Bengt Nilsson_2]

Ok, I think I got it.

# hwmgr -scan scsi

The "mt" command is now working, and also "vdump".

Question:
Which device file is now connected to my tape, if any?
If I do "# file /dev/*mt*" the info I get points to SCSI ID #0, which is wrong, should be 5. The same for /dev/tape and /dev/ntape.

So I am confused.
Any comments?

[Steven Schweda]

> My QLogic KZPBR-CY did not work, so I put
> it aside.

If that was actually a KZPBA-CB/CY (lots of
plug-in terminator resistor packs), then I'd
expect it to be a differential SCSI card
("HVD", in modern lingo), so I wouldn't
expect it to work with LVD/SE devices. (I
was once (2002) the only bidder ($8) on Ebay
for a "K2BPR-CX" card. Apparently, no one
else noticed that it was really a KZPBA-CX,
being offered by a seller who read it
wrong.)

> [...] Someone suggested it is necessary to
> REINSTALL V5.1B to detect and get new
> hardware support. Is this really true?

Tell that person to stand on one leg and
scream like a chicken, and then try:

dn_setup -boot

See, for example:

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1102238

> [...] the info I get points to SCSI ID #0
> [...]

What makes you think that any "0" there is
related to SCSI ID 0? (Is it true for the
disks in "/dev/disk"?)

[Rob Leadbeater · Accepted solution]

Can you post the output of "hwmgr view dev"

Cheers,
Rob

[Bengt Nilsson_2]

# hwmgr view dev
HWID: Device Name Mfg Model Location
------------------------------------------------------------------------------
3: /dev/dmapi/dmapi
4: /dev/scp_scsi
5: /dev/kevm
33: /dev/disk/floppy0c 3.5in floppy fdi0-unit-0
53: /dev/disk/cdrom0c SAMSUNG CD-ROM SC-148P bus-1-targ-0-lun-0
54: /dev/disk/dsk0c MAXTOR ATLAS10K5_300WLS bus-2-targ-0-lun-0
55: /dev/random
56: /dev/urandom
57: /dev/ntape/tape0 HP C7438A bus-2-targ-5-lun-0
92: /dev/disk/dsk1c COMPAQ BD01874554 bus-2-targ-1-lun-0

Ok, I see now that I did not realize the change in device files between 4.0F and 5.1B. "# file /dev/...." does not show the same info as in 4.0F, if any useful at all.
So I will RTFM and study hwmgr instead.

I think all is covered now.

I learnt (by myself actually) not to mix LVD/SE and LVD devices on the same bus;
I learnt not to assume ANY similarities between device management in 4.0F and 5.1B;

Thanks a lot!

[Bengt Nilsson_2]

And thanks Steven for informing me about the truth about my KZPBA-CB/CY, it explains a lot.
Closing this thread successfully.

[Rob Leadbeater]

Hi,

> I learnt not to assume ANY similarities
> between device management in 4.0F and 5.1B

Device management was one of the major changes between 4.0x and 5.x.

Glad to hear that you figured it out.

Remember to say thanks by assigning points.

Cheers,

Rob