Operating System - Tru64 Unix
1827853 Members
1596 Online
109969 Solutions
New Discussion

Re: mkfdmn problem

 
Amanda Deer
Frequent Advisor

mkfdmn problem

I am trying to create a domain on a new disk. I can disklabel it ok, but when I try to make the domain it just sits there doing nothing. Any ideas?
23 REPLIES 23
Ivan Ferreira
Honored Contributor

Re: mkfdmn problem

What is the command that you are using?

What is the output, or just hang?

Can you sucessfully dd to the disk to read and write some data?
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

Command is mkfdmn /dev/disk/dsk30c test

and it just hangs.
Michael Schulte zur Sur
Honored Contributor

Re: mkfdmn problem

Hi,

can you post
disklabel -r dsk30
file /dev/rdisk/dsk30c
?
Any errors in binary errorlog?

greetings,

Michael
Ivan Ferreira
Honored Contributor

Re: mkfdmn problem

Try this and verify that the command does not hangs. If the command hangs, there may be a hardware problem:

dd if=/dev/disk/dsk30c of=/dev/null bs=1024 count=100000

dd if=/dev/zero of=/dev/rdisk/dsk30c bs=1024 count=100000

Is this a local (DAS) disk or a SAN (or SCSI storage) disk?
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

dcs1a # file /dev/rdisk/dsk30c
/dev/rdisk/dsk30c: character special (19/699) SCSI #15 "HSZ70" disk #14 (SC
SI ID #2) (SCSI LUN #3) errors = 4/0

dcs1a # disklabel -r dsk30
# /dev/rdisk/dsk30c:
type: SCSI
disk: HSZ70
label:
flags:
bytes/sector: 512
sectors/track: 254
tracks/cylinder: 20
sectors/cylinder: 5080
cylinders: 41986
sectors/unit: 213291762
rpm: 3600
interleave: 1
trackskew: 7
cylinderskew: 26
headswitch: 0 # milliseconds
track-to-track seek: 0 # milliseconds
drivedata: 0

8 partitions:
# size offset fstype fsize bsize cpg # ~Cyl values
a: 131072 0 unused 0 0 # 0 - 25*
b: 262144 131072 unused 0 0 # 25*- 77*
c: 213291762 0 unused 0 0 # 0 - 41986*
d: 0 0 unused 0 0 # 0 - 0
e: 0 0 unused 0 0 # 0 - 0
f: 0 0 unused 0 0 # 0 - 0
g: 106449273 393216 unused 0 0 # 77*- 21031*
h: 106449273 106842489 unused 0 0 # 21031*- 41986*


Output from dd - the first command came back ok, and the second command returned this message.

dcs1a # dd if=/dev/zero of=/dev/rdisk/dsk30c bs=1024 count=100000
dd write error: Read-only file system
1+0 records in
0+0 records out
Venkatesh BL
Honored Contributor

Re: mkfdmn problem

Have you 'write-protect'ed the disk by any chance?
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

Not that I'm aware of.
Venkatesh BL
Honored Contributor

Re: mkfdmn problem

Because, the error suggests that the disk is 'read-only' (Note that the first command reads from the disk and the second command tries to write to the disk).

I would suggest that you check if the disk was configured as read-only on the controller side.
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

Logged onto the controller and looked at the unit. It says NOWRITE_PROTECT.
Vladimir Fabecic
Honored Contributor

Re: mkfdmn problem

Please post output of:
# ls -al /dev/rdisk/dsk30*
and
# ls -al /dev/disk/dsk30*
What is the type of controller?
In vino veritas, in VMS cluster
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

dcs1a # ls -al /dev/rdisk/dsk30*
crw------- 1 root system 19,695 Sep 14 10:56 /dev/rdisk/dsk30a
crw------- 1 root system 19,697 Sep 14 10:56 /dev/rdisk/dsk30b
crw------- 1 root system 19,699 Dec 12 09:48 /dev/rdisk/dsk30c
crw------- 1 root system 19,701 Sep 14 10:56 /dev/rdisk/dsk30d
crw------- 1 root system 19,703 Sep 14 10:56 /dev/rdisk/dsk30e
crw------- 1 root system 19,705 Sep 14 10:56 /dev/rdisk/dsk30f
crw------- 1 root system 19,707 Sep 14 10:56 /dev/rdisk/dsk30g
crw------- 1 root system 19,709 Sep 14 10:56 /dev/rdisk/dsk30h
dcs1a # ls -al /dev/disk/dsk30*
brw------- 1 root system 19,694 Sep 14 10:56 /dev/disk/dsk30a
brw------- 1 root system 19,696 Sep 14 10:56 /dev/disk/dsk30b
brw------- 1 root system 19,698 Sep 14 10:56 /dev/disk/dsk30c
brw------- 1 root system 19,700 Sep 14 10:56 /dev/disk/dsk30d
brw------- 1 root system 19,702 Sep 14 10:56 /dev/disk/dsk30e
brw------- 1 root system 19,704 Sep 14 10:56 /dev/disk/dsk30f
brw------- 1 root system 19,706 Sep 14 10:56 /dev/disk/dsk30g
brw------- 1 root system 19,708 Sep 14 10:56 /dev/disk/dsk30h


Controller is HSZ70.
Michael Schulte zur Sur
Honored Contributor

Re: mkfdmn problem

BL,

you can not write to the 16 first sectors.
They are write protected because there are crucial disk informations on it.
You will have to skip the first 16 sectors.
dd if=/dev/zero of=/dev/rdisk/dsk30c os=16 count=100000
213291762 sectors is a strange size. This makes it roughly 100GB. I can't remember such a disk. How did you label the disk?
Maybe we will have to redo it.

greetings,

Michael
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

disklabel -wr dsk30
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

I have tried disklabel -z dsk30 and then relabelling but still can't create the domain.
Michael Schulte zur Sur
Honored Contributor

Re: mkfdmn problem

Can you give us the disk model so we can
verify that it is supported by the HSZ70?
Please also post os version and patch kit version.
The disklabel looks the same as before?

Michael
Vladimir Fabecic
Honored Contributor

Re: mkfdmn problem

It is possible that it is a hardware problem, but not sure.
Can you do:
# newfs /dev/disk/dsk30c
and post the message?
In vino veritas, in VMS cluster
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

Ouput from newfs -

Warning: 2198 sector(s) in last cylinder unallocated
/dev/rdisk/dsk30c: 213291762 sectors in 41987 cylinders of 20 tracks, 254 s
ectors
104146.4MB in 2625 cyl groups (16 c/g, 39.69MB/g, 9536 i/g)
super-block backups (for fsck -b #) at:
32, 81568, 163104, 244640, 326176, 407712, 489248, 570784,
652320, 733856, 815392, 896928, 978464, 1060000, 1141536, 1223072,
1304608, 1386144, 1467680, 1549216, 1630752, 1712288, 1793824, 1875360,
1956896, 2038432, 2119968, 2201504, 2283040, 2364576, 2446112, 2527648,
2600992, 2682528, 2764064, 2845600, 2927136, 3008672, 3090208, 3171744,
3253280, 3334816, 3416352, 3497888, 3579424, 3660960, 3742496, 3824032,
3905568, 3987104, 4068640, 4150176, 4231712, 4313248, 4394784, 4476320,
4557856, 4639392, 4720928, 4802464, 4884000, 4965536, 5047072, 5128608,
5201952, 5283488, 5365024, 5446560, 5528096, 5609632, 5691168, 5772704,
5854240, 5935776, 6017312, 6098848, 6180384, 6261920, 6343456, 6424992,
6506528, 6588064, 6669600, 6751136, 6832672, 6914208, 6995744, 7077280,
7158816, 7240352, 7321888, 7403424, 7484960, 7566496, 7648032, 7729568,
7802912, 7884448, 7965984, 8047520, 8129056, 8210592, 8292128, 8373664,
8455200, 8536736, 8618272, 8699808, 8781344, 8862880, 8944416, 9025952,
9107488, 9189024, 9270560, 9352096, 9433632, 9515168, 9596704, 9678240,
Venkatesh BL
Honored Contributor

Re: mkfdmn problem

well...well...well...looks like the disk is 'writable' after all!

newfs succeeds (does the disklabel output show 4.2 BSD for dsk30c?) - that hints that mkfdmn is the culprit.

Vladimir Fabecic
Honored Contributor

Re: mkfdmn problem

Looks like you can create ufs file system with no problem. If you can mount it, that means it is probably not a hardware problem (still not sure).
Can you post output of:
# ls -alR /etc/fdmns
In vino veritas, in VMS cluster
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

8 partitions:
# size offset fstype fsize bsize cpg # ~Cyl values
a: 131072 0 unused 0 0 # 0 - 25*
b: 262144 131072 unused 0 0 # 25*- 77*
c: 213291762 0 4.2BSD 1024 8192 16 # 0 - 41986*
d: 0 0 unused 0 0 # 0 - 0
e: 0 0 unused 0 0 # 0 - 0
f: 0 0 unused 0 0 # 0 - 0
g: 106449273 393216 unused 0 0 # 77*- 21031*
h: 106449273 106842489 unused 0 0 # 21031*- 41986*

I am raising this with HP now. Thanks for your help.
Vladimir Fabecic
Honored Contributor

Re: mkfdmn problem

Please check that maybe you have another domain using dsk30.
# ls -alR /etc/fdmns | grep dsk30
# ls -alR /etc/fdmns | grep test
In vino veritas, in VMS cluster
Amanda Deer
Frequent Advisor

Re: mkfdmn problem

Already checked that and no they're not being used already.
Michael Schulte zur Sur
Honored Contributor

Re: mkfdmn problem

mkfdmn would also have complained about a partition already marked for use.
Can you please give us the disk model?

thanks,

Michael