ibm scsi disk install ?

[TIM HUGHES_2]

have disks (8x36 IBM Ultrastar) installed
and recognised by the host, the device files
are/have been created (b & c), but i can not
access the drives - in sam it shows the drives
as having 0 (zero) MB???:

stm shows 8 entries like this:

=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=-+-=

-- Information Tool Log for SCSI Disk on path 0/7/0/0.0.0 --

Log creation time: Wed Jul 9 14:18:05 2003

Hardware path: 0/7/0/0.0.0


The selected device is not supported by this tool.
Inaccurate information could be displayed by the tool.

Product Id: IC35L036UCDY10-0 Vendor: IBM
Device Type: SCSI Disk Firmware Rev: S23C
Device Qualifier: IBMIC35L036UCDY10-0 Logical Unit: 0
Serial Number: E3L08MPB
Capacity (M Byte): 35003.59
Block Size: 512
Max Block Address: 71687339
Error Logs
Read Errors: 0 Buffer Overruns: N/A
Read Reverse Errors: N/A Buffer Underruns: N/A
Write Errors: 0 Non-Medium Errors: 0
Verify Errors: 0


print manifest shows 8 of these:
IBM IC35L036UCDY10-0 floppy 0/7/0/1.3.0 SCSI C1010 Ultra160 Wide LVD A6829-60101

ioscan shows:
# ioscan -n -k -f /dev/rdsk/c4t0d0
Class I H/W Path Driver S/W State H/W Type Description
=====================================================================
disk 7 0/7/0/0.0.0 sdisk CLAIMED DEVICE IBM IC35L036UCDY10-0
/dev/dsk/c4t0d0 /dev/rdsk/c4t0d0

sam show them as 0 MB:
# cat /tmp/sam.Log
Hardware Number Volume Total
Path of Paths Use Group Mbytes Description
0/0/1/0.1.0 1 Unused -- 17501 IBM DPSS-318350N
0/0/1/0.2.0 1 Unused -- 17501 IBM DPSS-318350N
0/0/1/0.3.0 1 Unused -- 17501 IBM DPSS-318350N
0/0/1/0.4.0 1 Unused -- 17501 IBM DPSS-318350N
0/0/1/1.0.0 1 LVM vg01 8683 SEAGATE ST39204LC
0/0/2/0.0.0 1 LVM vg00 8683 SEAGATE ST39204LC
0/0/2/1.2.0 1 CDFS -- 208 HP DVD-ROM SCSI Drive
0/7/0/0.0.0 1 Unused -- 0 IBM IC35L036UCDY10-0
0/7/0/0.1.0 1 Unused -- 0 IBM IC35L036UCDY10-0
0/7/0/0.2.0 1 Unused -- 0 IBM IC35L036UCDY10-0
0/7/0/0.3.0 1 Unused -- 0 IBM IC35L036UCDY10-0
0/7/0/1.0.0 1 Unused -- 0 IBM IC35L036UCDY10-0
0/7/0/1.1.0 1 Unused -- 0 IBM IC35L036UCDY10-0
0/7/0/1.2.0 1 Unused -- 0 IBM IC35L036UCDY10-0
0/7/0/1.3.0 1 Unused -- 0 IBM IC35L036UCDY10-0


any thoughts?

thxx - tim

[John Poff]

Hi,

What does 'diskinfo' report? My guess is that it is a firmware issue with the disks.

JP

[A. Clay Stephenson]

I would forget about using sam. You have the device nodes so I would do a pvcreate using the command line --- like real Sys Admins.

e.g. pvcreate /dev/rdsk/c4t0d0

You may also have to specify a -s (size) paramter if the OS cannot determine the disk size but I'm betting that it can.

You will then proceed with vgcreate.

Man pvcreate and vgcreate for details.

You can then finish the job using SAM or lvcreate and newfs.

[Paul Sperry]

mkdir /dev/vg#1
mknod /dev/vg#1/group c 64 0x010000
pvcreate /dev/rdsk/c#t#d#
vgcreate /dev/vg#1 /dev/rdsk/c#t#d#
lvcreate -L -i 3 -I 64 -n datavol /dev/vg#1
newfs -F vxfs -o largefiles /dev/vg#1/datavol

or something like that :)

[TIM HUGHES_2]

hmmmm -

diskinfo shows:

# diskinfo -v /dev/rdsk/c4t0d0
SCSI describe of /dev/rdsk/c4t0d0:
vendor: IBM
product id: IC35L036UCDY10-0
type: direct access
size: 0 Kbytes
bytes per sector: 0
rev level: S23C
blocks per disk: 0
ISO version: 0
ECMA version: 0
ANSI version: 3
removable media: no
response format: 2
(Additional inquiry bytes: (...clipped...)

what i really want to use the disk for is a
raw db device?

if i try to put an fs or vg on it it either
fails?

[TIM HUGHES_2]

thanks for the responses! - still apparently no good?

i tried creating the pvg and vg to no avail?

# pvcreate /dev/rdsk/c4t0d0
# mkdir /dev/vgdb01
# mknod /dev/vgdb01/group c 64 0x050000
# vgcreate /dev/vgdb01 /dev/dsk/c4t0d0

vgcreate: Physical Volume "/dev/dsk/c4t0d0" contains no LVM information

and no vg is created...?

maybe a firmware prob like JP suggests?
one more shot with a -s on the pvcreate

[Steven E. Protter]

I think you have a SCSI ID problem.

You have 8 disks but 10 show up on I/O Scan.

Sounds like Mr. Scott needs to fix the transporter.

While attaching a HP6000 SE disk array to a D320 that will soon be in my basement, I accidentally reconnected a connector that should have been left alone.

Suddenly there were SCSI timeouts and more disks showed up on ioscan than were actually there.

After frustration, I remembered plugging in the cable, which I thought I had disconnected, and re-disconnected it. Everything was fine.

I don't now how the ID is set on your drives, but there is something in there confusing your server.

SEP

[TIM HUGHES_2]

i think the scsi ids are okay ioscan show the 8 IBM drives okay - there are several other drives attached to the system that continue to operate correctly...???

pvcreate fails with a '-s' option:

# pvcreate -s 512 /dev/rdsk/c4t0d0
pvcreate: writing LVM record: Invalid argument

rats...

ioscan show the devices on these two controllers correctly?
0/7/0/0 ext_bus SCSI C1010 Ultra160 Wide LVD A6829-60101
0/7/0/0.0 target
0/7/0/0.0.0 disk IBM IC35L036UCDY10-0
0/7/0/0.1 target
0/7/0/0.1.0 disk IBM IC35L036UCDY10-0
0/7/0/0.2 target
0/7/0/0.2.0 disk IBM IC35L036UCDY10-0
0/7/0/0.3 target
0/7/0/0.3.0 disk IBM IC35L036UCDY10-0
0/7/0/0.7 target
0/7/0/0.7.0 ctl Initiator
0/7/0/1 ext_bus SCSI C1010 Ultra160 Wide LVD A6829-60101
0/7/0/1.0 target
0/7/0/1.0.0 disk IBM IC35L036UCDY10-0
0/7/0/1.1 target
0/7/0/1.1.0 disk IBM IC35L036UCDY10-0
0/7/0/1.2 target
0/7/0/1.2.0 disk IBM IC35L036UCDY10-0
0/7/0/1.3 target
0/7/0/1.3.0 disk IBM IC35L036UCDY10-0
0/7/0/1.7 target
0/7/0/1.7.0 ctl Initiator

[A. Clay Stephenson]

It appears that your SCSI ID's are ok (0-3) on two different SCSI buses. It's now time to check some more fundamental things:

1) Termination in exactly two places on each bus -- and must be on the ends of the bus. Check to make sure that termination is not enabled on anything except the controller and the last drive on the bus (if not terminated with a dedicated terminator on the end).

2) At least one device enabled to supply terminator power. -- jumper setting

3) It's possible that your drives have not been enabled to "auto-spin" (i.e. spin up on power on automatically). This is a jumper setting.

I also noticed that your initial ioscan used the -k option; I would really like to see an ioscan -fn that ignores what the kernel "thinks" is the view of the devices.

I would also try a direct i/o of the disk using dd.

e.g. dd if=/dev/zero bs=64k count=1000 of=/dev/rdsk/c4t0d0.

[TIM HUGHES_2]

thanks again for sticking with me on this one.

i'm not much of an hp-hpux admin but slowly
getting better, i appreciate your help.

i've attached a zipped txt output file from:
(edited for pertinent info [imho])

ioscan -fn
cstm infolog

i'm going to look at the hardware settings
on the disks (terminations and jumpers)

do i need an entry in the /etc/diskinfo file
for these disks? for them to recognised correctly?

thxx again - tim

[TIM HUGHES_2]

otay...

drives sit in an array enclosure -sdsi id is determined by the array - not hard jumpered on the drives...

as a matter of fact the drives are not jumpered
for any option.

they are Ultra 320s (LVDs) actually now
the UltraStar are joint venture between hitachi and ibm... (model is Ultrastar
146Z10 - 36Gb model - 80 pin interface)
so they can't/don't even supply termination
power (it must be supplied by the host/array)
which it is...

they are not jumpered for autospin disable(
on the 80 pin this jumper diables autospin)
or delay spinup...

the interesting thing i have seen is that print_manifest and stm think they are removable
media - and i am wondering if this is the cause of the system thinking they are 0 Kb?
print_manifests actually lists them as floppy
drives - and stm when you try to do anything with them besides interrogate info, it requests that media be loaded (repeatedly...)

is there something i can do about that? any ideas/direction/information is always appreciated.

thxx - tim