Tru64 Unix Upgrade problem

[Kafsat Taiyus]

Hi,

I have tried to upgrade a MCALPHA4000 server from Try64 Unix v5.1 to v5.1b. I have updated all the firm ware before starting the upgrade.

During upgrade it gave me these wornings.
"The following software may require reinstallation after the Update Install is Complete.

Storageworks RAID Array 200 Management Utility V1.2.0
Storage works SWXCR Utily from Digital Unix V1.2.3"

I ignored the worning and continued with the update.

At the end of update when the sytem is re-booted it can not see any of the disks. It boots into single user mode. It mounts only "/" file system and fails to mount any other file system.

If I try to mount file systems manually it gives error like "/dev/disk/dsk0 is invalid device or cannot be opend"

What can be done to fix the system?

Thank you for your help.

Regards
Kafsat

[Michael Schulte zur Sur]

Hi,

these warnings are nothing to worry about. It just says that if the software stops working, reinstall it. But so far I haven't had to do it. What is the firmware version you installed and what kind if mylex controller have you got? May I assume, it controls the boot disk?
Please post
df -k
hwmgr -show scsi

greetings,

Michael

[Kafsat Taiyus]

Michael,
Thank you for such quick response.

df -k
FileSystem 1020-Block Used Available Capacity Mounted on
root_domain#root 393216 109066 274284 24% /

hwmger -show scci

53 0 cdrom none 0 1 (null) [0/4/0]
56 1 tape none 0 1 (null) [0/5/0
57 2 tape none 0 1 (null)

Thank you.

[Michael Schulte zur Sur]

Ok,

there are a few things that have been obsoleted. Have you checked that the hardware in your box is still supported?
What firmware level have you updated to?
Can you post
showfdmn -k root_domain
file /dev/rdisk/dsk*

Michael

[Vladimir Fabecic]

Did you try
>>> boot -file genvmunix

[Kafsat Taiyus]

Michael,

Milex Card DAC960.
AlphaBIOS V5.70
SRM flash V6.1-1

I did not check the hardware.

showfdmn -k root_domain
3e2b77d600020f22 Mon Jan 20 12:15:18: 2003 512 4 root_domain

Vol 1K-BLKS Free %used cmode Rblks Wblks Vol Name
12 393216 274248 30% on 128 128 /dev/disk/dsk0a

I could not run file /dev/rdisk/dsk* command file is not available.

I look forward to your response.

Thank you

[Michael Schulte zur Sur]

Kafsat,

accroding to my information there is no SRM version 6.1
It is important to know what kind of controller the mylex is. KZPSC is not supported anymore.
The error message is strange because root_domain is using dsk0a.

You can check dsk0 with
dd if=/dev/rdisk/dsk0c of=/dev/null bs=1024k

Michael

[Ivan Ferreira]

I also think that you should try booting the genvmunix kernel to redetect the hardware. It seems that you loosed the special file for the disks.

>>> boot -fi genvmunix -s

Recompile the kernel with doconfig.
Copy the new kernel to /

Reboot the server.

Device special files are created with dsfmgr -k or dsfmgr -K, try using those commans too.

[Johan Brusche]

An AlphaServer, like the 4100, highest SRM is V6.1-1, loaded from firmware-CD V6.7 or higher.

__ Johan.

[Michael Schulte zur Sur]

Johan,

I just booted from v6.9.
srm-flash shows up as v6.0-4

Michael

[Kafsat Taiyus]

Thank you everybody

I booted the machine using
boot -file genvmunix same thing


I tried to create device special files using dsfmgr -K it gives follwoing errors

dsfmgr -K
dsfmgr: ERROR: Cannot open and read file /etc/dfsc.dat: Bad status file data/format

dsfmgr: ERROR: Cannot open and read file /etc/dfsl.dat: Bad status file data/format

Looking forward to your responses.

Regards
Kafsat

[Kafsat Taiyus]

I have got around this problem. I run following commands

dn_setup -init
dsfmgr -K

Now the disk is visible and the installation is conitnuing.

Thank you everybody for your help.

Regards
Kafsat

[Johan Brusche]

root cause.......

inbetween V5.1A and V5.1B, the format of the dfs[l|c].dat files changed from something like:

I: dsk 5
I: tape 10
I: floppy 1
I: cdrom 1
# 1.0 V5.1A (Rev. 1885) Mon Feb 7 14:51:11 2005
# ldev cdev hw_id bn_id dn_id typ legacy name
A: 0 1300011 39 7 0 b "" /dev/disk/ cdrom 0 a

to:

I: floppy 2
I: cdrom 4
I: dsk 78
I: scp 6
I: mc 9
I: tape 19
V: 1.1 V5.1B (Rev. 2650) Wed Jul 20 11:03:55 2005
# ldev cdev hw_id bn_id dn_id typ legacy name
A: 0 13002f1 287 72 0 c "" /dev/changer/ mc 5 ""


During the upgrade the conversion is performed by /etc/dn_fix_dat.sh, but in your case the exit status must have been "not succesfull"

You could have run this script manually to check what it said, before doing the dn_setup -init .

You should find traces of what happened in /var/adm/smlogs/update.log

__ Johan.