Swap not detected

[tru64 unix_1]

vm_swap_init: warning /dev/disk/dsk0d swap device not found
vm_swap_init: swap is re-set to lazy (over commitment) mode

I can not boot one DS10, at the begining does not detect osf_boot, when I could boot it with the CDROM and mount it, I find out that osf_boot, genvmunix and vmunix where missing. I copied missing files first from CDROM, then I have tried from other machines but still not boot. I have read that swap possibly means there is a problem related with LUN's. I attached the information collected during booting time, swapon and hwmgr -s scsi. Can anyone give some tips?

[Michael Schulte zur Sur]

Hi,

check /etc/sysconfigtag for the swap definition and the disklabel of dsk0 whether partition d is labeled for swap.

greetings,

Michael

[Ivan Ferreira]

These files that you copied maybe are not the only files that where deleted. I won't trust this system in the state that is right now. If you have a backup, restore the backup. Your disk is there, but maybe the hardware database or devices files of the system is corrupt or missing, so it cannot find the swap disk.

You may try booting in single user with the generic kernel:

boot -fi genvmunix -fl s

[tru64 unix_1]

I have tried a Restore of the disklabel but still gives swap device not found. I do not have a backup available, do you know how can I verify the HW database?

[Mark Poeschl_2]

Booting as Ivan suggested from the generic kernel and into single mode is what you need to try next. From there you MAY be able to reconstruct a usable system, but as pointed out in previous posts, it's likely that other things have been deleted and without a backup you're stuck with re-installing from scratch. If you CAN'T get to single user mode using the generic kernel, you're almost certainly facing a re-installaing. There are WAY too many things that could be causing that - and a missing swap device isn't even one of them. :-(

[tru64 unix_1]

I have tried single user mode but the result is the same.
:(
I would try to avoid reinstallation until get the DB information need it but it seems there is no other option.

Thanks for your help :(

[Johan Brusche]

Are you sure dsk0d is really your swap, and not the default dsk0b ?

Anyway when booted from CD, you should be able to do:

cd /etc/fdmns
mkdir DBdomain
cd DBdomain
ln -s /dev/dsk1c
showfsets DBdomain

mount DBdomain#dbset /mnt
dn_setup install_tape

vdump -0f /dev/ntape/tape0c /mnt

Rgds,
___ Johan.

[Mark Poeschl_2]

Is the 'DB information' on your second drive (dsk1) or is it also on dsk0? If you re-use dsk0 as your OS disk, everything on dsk1 will still be intact. The re-installation process only overwrites those partitions used for /, /usr, /var, and swap.

[tru64 unix_1]

I did not install the system and I do not why they choose b and d as swap, at the beginning I was also uncertain about this but I compare with another system installed in similar configuration.
Thanks for your help.

[tru64 unix_1]

I have to reinstall the system from scratch.
Thanks for your effort to try to avoid this situation.
Now it is up again.