DLT7000 tape drive problem

[Ronelle van Niekerk]

Hi guys,

We've connected a compaq mini-library to an L-Class and, after some messing around with the terminators, we managed to get the server to see the HVD card and the autochanger but it won't see the 2 DLT drives.

We've installed all the drivers - we think - but still no luck.

Any ideas?

-Ronelle

[A. Clay Stephenson]

Do an ioscan -fn and post the results.

[Ronelle van Niekerk]

Here's the ioscan output:

ext_bus 5 0/3/0/0 c720 CLAIMED INTERFACE SCSI C87x Fast Wi
de Differential
target 7 0/3/0/0.1 tgt CLAIMED DEVICE
autoch 0 0/3/0/0.1.0 schgr CLAIMED DEVICE DEC TL800
(C) DEC
/dev/rac/c5t1d0
target 8 0/3/0/0.7 tgt CLAIMED DEVICE
ctl 4 0/3/0/0.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c5t7d0

[A. Clay Stephenson]

This can't be the complete list. Don't edit it and don't apply any -C xxxx ioscan filter just do an ioscan -fn.

[Ronelle van Niekerk]

OK then, here's the entire ioscan:

Class I H/W Path Driver S/W State H/W Type Description
======================================================================
lan 0 0/0/0/0 btlan CLAIMED INTERFACE HP PCI 10/100Base-TX C
ore
/dev/diag/lan0 /dev/ether0 /dev/lan0
ext_bus 0 0/0/1/0 c720 CLAIMED INTERFACE SCSI C896 Fast Wide LV
D
ctl 0 0/0/1/0.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c0t7d0
ext_bus 1 0/0/1/1 c720 CLAIMED INTERFACE SCSI C896 Ultra Wide S
ingle-Ended
disk 0 0/0/1/1.2.0 sdisk CLAIMED DEVICE SEAGATE ST318203LC
/dev/dsk/c1t2d0 /dev/rdsk/c1t2d0
ctl 1 0/0/1/1.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c1t7d0
ext_bus 2 0/0/2/0 c720 CLAIMED INTERFACE SCSI C87x Ultra Wide S
ingle-Ended
disk 1 0/0/2/0.2.0 sdisk CLAIMED DEVICE SEAGATE ST318203LC
/dev/dsk/c2t2d0 /dev/rdsk/c2t2d0
ctl 2 0/0/2/0.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c2t7d0
ext_bus 3 0/0/2/1 c720 CLAIMED INTERFACE SCSI C87x Ultra Wide S
ingle-Ended
disk 2 0/0/2/1.2.0 sdisk CLAIMED DEVICE HP DVD-ROM 304
/dev/dsk/c3t2d0 /dev/rdsk/c3t2d0
ctl 3 0/0/2/1.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c3t7d0
tty 0 0/0/4/0 asio0 CLAIMED INTERFACE PCI Serial (103c1048)
/dev/GSPdiag1 /dev/mux0 /dev/tty0p1
/dev/diag/mux0 /dev/tty0p0 /dev/tty0p2
tty 1 0/0/5/0 asio0 CLAIMED INTERFACE PCI Serial (103c1048)
/dev/GSPdiag2 /dev/mux1
/dev/diag/mux1 /dev/tty1p1
ext_bus 5 0/3/0/0 c720 CLAIMED INTERFACE SCSI C87x Fast Wide Di
fferential
autoch 0 0/3/0/0.1.0 schgr CLAIMED DEVICE DEC TL800 (C) D
EC
/dev/rac/c5t1d0
ctl 4 0/3/0/0.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c5t7d0
fc 0 0/4/0/0 td CLAIMED INTERFACE HP Tachyon TL/TS Fibre
Channel Mass Storage Adapter
/dev/td0
fcp 0 0/4/0/0.8 fcp CLAIMED INTERFACE FCP Protocol Adapter
ext_bus 4 0/4/0/0.8.0.255.0 fcpdev CLAIMED INTERFACE FCP Device Interface
disk 3 0/4/0/0.8.0.255.0.3.0 sdisk CLAIMED DEVICE DGC C340
0WDR5
/dev/dsk/c4t3d0 /dev/rdsk/c4t3d0
disk 4 0/4/0/0.8.0.255.0.3.1 sdisk CLAIMED DEVICE DGC C340
0WDR5
/dev/dsk/c4t3d1 /dev/rdsk/c4t3d1
disk 5 0/4/0/0.8.0.255.0.3.2 sdisk CLAIMED DEVICE DGC C340
0WDR5
/dev/dsk/c4t3d2 /dev/rdsk/c4t3d2
disk 6 0/4/0/0.8.0.255.0.3.3 sdisk CLAIMED DEVICE DGC C340
0WDR5
/dev/dsk/c4t3d3 /dev/rdsk/c4t3d3
disk 7 0/4/0/0.8.0.255.0.3.4 sdisk CLAIMED DEVICE DGC C340
0WDR5
/dev/dsk/c4t3d4 /dev/rdsk/c4t3d4
disk 8 0/4/0/0.8.0.255.0.3.5 sdisk CLAIMED DEVICE DGC C340
0WDR5
/dev/dsk/c4t3d5 /dev/rdsk/c4t3d5

[Chris Vail]

I suggest you replace the cables with those that you know beyond any doubt are good. This ioscan should've found the drives, but obviously it hasn't. Has the sytem been rebooted since installing this system. In theory, this should make no difference, but t'wouldn't hurt to try.

Also make sure that you're using the right kind of DLT's. You could be using differential drives on a single ended channel or vice versa. They use the same cabling. Check the symbols on the connectors. You may need to change out the cables, SCSI controller, or the drives themselves.

Good luck
Chris

[Ronelle van Niekerk]

Thanks chris. But doesn't the fact that the ioscan found the autochanger mean the scsi card and cable are fine. (Installed an HVD card specially for this library) and the system has been rebooted several time since we installed the card this morning (for kernel drivers and such.)

[Steven E. Protter]

insf -e

Seems that everything is okay in SCSI-land.

The command above might let you see the two drives.

I'm wondering if this DEC tape drive is certified to work with this hardware.

There has to be a more detailed install document on this product.

Perhaps a document search in the compaq section of itrc.

SEP

[Ronelle van Niekerk]

Thanks Steven

The insf -e unfortunately did not help.

I'll see if I can find more documentation on the library.

[Ronelle van Niekerk]

Went to the website and foudn that the supported operating systems are:

Windows NT 4.0
Windows 2000
Novell NetWare 4.1, 4.2, 5.0
Tru64Unix 4.0e, 4.0f, 5.0a
Open VMS 6.2-1h3, 7.1-2, 7.2-1

So I guess we're going to have to find another library somewhere for our testing.

[A. Clay Stephenson · Accepted solution]

I wouldn't give up too quickly. HP-UX (as are most UNIX's) is very tolerent of SCSI devices. I've seen only a very few tape drives that would not work. Because the drives are not even seen suggests that you have a cabling, SCSI ID, or termination problem.

First, note that more than 1 DLT7000/8000 per bus is not supported. A robot and drive are allowed to share a common bus iff (if and only if) they are both of compatible architecture. I've seen a few libraries where the robot was a SE-SCSI device while the drives were HVD SCSI. The most likely problem is that the drives are assigned a common SCSI ID; if originally these drives were on separate buses that would not have been a problem but now (if as I suspect they are daisy-chained on the same bus) you have an instant address conflict. Make sure that this bus is terminated in EXACTLY two places -- on the ends of the bus. You should also make sure that at least one device on the bus is supplying termination power. You also need to find documentation about the drives and robotics and make certain of the SCSI standards used by each. Also, make sure that if this HVD SCSI that even the terminators employed are HVD devices. Just because the connectors/cables physically fit means nothing.

[Ronelle van Niekerk]

Clay,

Your post got me playing around with the terminators on the library and I came up with one tape drive!

ioscan now looks like this:

ext_bus 5 0/3/0/0 c720 CLAIMED INTERFACE SCSI C87x Fast Wide Di
fferential
autoch 0 0/3/0/0.1.0 schgr CLAIMED DEVICE DEC TL800 (C) D
EC
/dev/rac/c5t1d0
tape 0 0/3/0/0.2.0 stape CLAIMED DEVICE DEC TZ89 (C) D
EC
ctl 4 0/3/0/0.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c5t7d0



The back of the library now looks as such:

LIB1 | DLT1 | DLT2
| |
B C | C S | T T
|_____|

B = blank
C = cable
S = server
T = terminator

The library menu tells me that DLT1 is SCSI ID 2 and DLT2 is SCSI ID 3.

I'm happy to go forward with just the one tape drive but would just like to know if this set up is safe.

[Ronelle van Niekerk]

It's looking even better now:

ext_bus 5 0/3/0/0 c720 CLAIMED INTERFACE SCSI C87x Fast Wi
de Differential
target 7 0/3/0/0.1 tgt CLAIMED DEVICE
autoch 0 0/3/0/0.1.0 schgr CLAIMED DEVICE DEC TL800
(C) DEC
/dev/rac/c5t1d0
target 10 0/3/0/0.2 tgt CLAIMED DEVICE
tape 0 0/3/0/0.2.0 stape CLAIMED DEVICE DEC TZ89
(C) DEC
/dev/rmt/0m /dev/rmt/c5t2d0BEST
/dev/rmt/0mb /dev/rmt/c5t2d0BESTb
/dev/rmt/0mn /dev/rmt/c5t2d0BESTn
/dev/rmt/0mnb /dev/rmt/c5t2d0BESTnb
target 8 0/3/0/0.7 tgt CLAIMED DEVICE
ctl 4 0/3/0/0.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c5t7d0

[A. Clay Stephenson]

If I can follow you diagram, it appears that you have 2 terminators on the same end of the bus.

Here is the ideal setup

Term --- Host SCSI Controller (ID7) ---- Library Robot (ID1) --- DLT1 (ID2) --- Term.

An optional configuration:
Term --- Host SCSI Controller (ID7) ---- Library Robot (ID1) --- DLT1 (ID2) --- DLT2 (ID3) --- Term.

Note that a terminator is typically already physically installed on the SCSI controller.
Depending upon the controller mode, some are
enabled via firmware commands and some are enabled using jumpers and still others are enabled by physically moving resistor packs from a set of dummy sockets to the "real" sockets. Make sure that the termination is enabled/configued on the controller itself. That takes care of one end, now you need to make certain that you have a HVD/FWD terminator on the other end of the SCSI bus. That is typically simply plugged on to the end of the chain but in a few cases you enable termination on a device by jumpers or software commands.

Note: The optional configuratiuon (2 drives on the same bus) works iff there is but one ACTIVE drive at any one time.

Depending upon the software used to access the library, you may need to setup a SCSI pass-thru device node rather than the schgr device node.

[Ronelle van Niekerk]

Thanks Clay.

I know I had two terminators on the one end but, as I said, I was playing around and that config just worked.

I took one of the terminators off this morning and it still works so I'll leave it like that for now.

[A. Clay Stephenson]

SCSI buses are rather funny beasts. Poorly terminated, over terminated, or non-terminated buses will in many cases work almost perfectly --- the very worst kind of problem to fight. If you have another HVD controller, DLT should go on its own bus but it should work in the daisy-chained configuration I showed you. In most cases, a single DLT on a bus well tremenously outperform two daisy-chained drives because 1 drive it all but the limit for a fully streaming DLT. If you try to use two, both will be starved for data and the drives will write/rewind/write in very small burst and the transfer rate will drop by a factor > 10.
It's fine to put both drives on the same bus as long as you don't spin 'em simultaneously.

[Ronelle van Niekerk]

Thanks for all your help Clay.

I think I'll just use it as is - it's just for my testing purposes for a week or two anyway and I don't think I'll be using more than one drive at a time.

[Albert Smith_1]

It's been a while since I worked with a DLT7000 library but here is a stupid question. Is your L class running 11i?

The reason I ask is that in most of the autochangers you need to install the past through driver (SPT) to run the robot the reason is that the robot shares a controller with one of the tape drives.

Here is the problem 11i does not support the pass through. You would need to manually control operation of one of the drives or the robot. From what I am seeing from your output you might be experincing that problem.

[Ronelle van Niekerk]

Ah, yes - the server is running 11i.

But luckily I will be downgrading to 11.00 for the actual testing so maybe it will work properly then - with the pass through driver.

[Steve Robinson_12]

Please Note the SPT driver was for devices using the tape2 (NIO) driver as you are using stape and have schgr driver in the kernel then from what I see you should be able to change the media using the mc command

for example to check slot status:

[root@hp] mc -p /dev/rac/cXtXdX -r DS
DT_slot_1 FULL
ST_slot_1 EMPTY
ST_slot_2 EMPTY
ST_slot_3 EMPTY
ST_slot_4 EMPTY
ST_slot_5 EMPTY
ST_slot_6 FULL

[Ronelle van Niekerk]

mc command came back with:

ERROR: 0x0 No Sense : 0x0 No additional sense information

[Steve Robinson_12]

From previous info:
autoch 0 0/3/0/0.1.0 schgr CLAIMED DEVICE DEC TL800 C) DEC /dev/rac/c5t1d0

So if you did a
#mc -p /dev/rac/c5t1d0 -r DS

and this returned a sense key 0x0 then we are not talking to this device correctly.

You could try

#ioscan -fnkH 0/3/0/
confirm the picker device is /dev/rac/c5t1d0

Then in /usr/contrib/src we have a scsi_io.c program we can use to probe the picker device, we do this like this;

#cd /usr/contrib/src
# ls scsi_io.c (confirm exists)
#cc -o check_picker scsi_io.c
once complied we can:
#./check_picker /dev/rac/c5t1d0

In my case using a DDS library I get;

# ./check_picker /dev/rac/c1t3d1
The thing claims to be: HP C1553A NS00 58

If you still get a scsi sense key errors then this suggest you still have a bus conectivity/termination issue or the DLT firmware is incompatable with HPUX (review your stape and scsi patches) at this point as the device is not supported on HPUX I guess you will have to make do with manually moving tapes.

[Ronelle van Niekerk]

I think I'm going to wait unti lI downgrade back to 11.00 and then try the library again.

Thanks for all your help.

[A. Clay Stephenson]

Before you do that, let's setup a scsi-passthru (sctl) device for the robot -- which is supported under 11.11. SCSI Pass-thru is generally my weapon of first chice rather than autochanger devices. Make sure that sctl driver is installed in the kernel.

1) lsdev and look for the sctl character device major number; it's typically 203 (dec).

2) mkdir /dev/robot
3) mknod /dev/robot/c5t1d0 c 203 0x051000
4) chmod 666 /dev/robot/c5t1d0

Now try loading tapes using this device.

[Ronelle van Niekerk]

Thanks Clay,

I may try that if I have time this week. If not, I will downgrade my system next week and try my luck on 11.00/