fail to boot HPUX

HPUX running in D-Class (L1000), pretty old HPUX version - hpux 11.00

Attempt 1 -- To boot from normal (primary)

Unable to boot - system complains failure
SYSTEM ALERT System Name : uninitialized DATE : 10/22/2011 Time : 03/41:12 Alert Level 15 = Fatal hardware or configuration problem prevents operation Reason for alert source : 8 = I/O Source Detail : 6 = disk Source ID :0 Problem Detail : 3 functional failure

Attempt 2 -- To boot from alternate disk
Booted and go to ISL prompt and try to startup in single user hpux -is
Error display complain on the Exec Failed and couldn't find the /stand/vmunix or /vmunix

:: This is to show that the kernel corrupted in the disk

Attempt 3 -- To boot from Recovery Tape
Able to boot by select recover shell (including loadfile vgcreate vgchange lvlnboot ...)
also frecover some LVM files from fbakcup tape and all the /etc/lvmtab /etc/fstab and /etc/mnttab
successfully loaded

When try to create the vg00, following error happen

lvlnboot -v
lvlnboot : Warning Can not determine all plysical volumes on which mirrores copes of the logical volume are located:
"/etc/lvmtab" is missing Physical Volumes.
Boot definitions for Volume Group /dev/vg00
Physical Volumes beloing in Root Volume Group :
/dev/dsk/c1t0d0 (0/0/1/1.0.0) -- Boot Disk
Unable to query the I/O interface : Parameter error
lnlnboot : unable to print reliable information

pvdisplay /dev/dsk/c1t0d0

-- Physical volumes --

PV Name /dev/dsk/c1t0d0
VG Name /dev/vg00
PV Status available
Allocatable yes
VGDA 2
CUR LV 8
PE Size (MBytes) 4
Total PE 4340
Free PE 18
Allocated PE 4322
Stale PE 0
IO Timeout (seconds) default

lvdisplay -v /dev/vg00/lvol1
-- Logical volumes ---

LV Name /dev/vg00/lvol1
VG Name /dev/vg00
LV Permission read/write
LV Status available/stale
Mirror copies 1
Consistency Recovery MWC
Schedule Parallel
LV Size (MBytes) 512
Current LE 128
Allocated PE 256
Stripes 0
Stripe Size (Kbytes) 0
Bad Block off
Allocation strict/contigous
IO Timeout (Seconds) default

----- Distribution of logical volume ----

PV Name LE on PV PE on LV
/dev/dsk/c1t0d0 128 128

Logical extents --

LE PV1 PE1 Status PV2 PE2 Status 2
00000 ??? 000000 stale /dev/dsk/c1t0d0 00000 current
..

From the above, the harddisk is corrupted. So I am thinking to break the mirror so that I can use one of the usable disk. (This server currently is only use for reference purposes only). Thus, no intention to replace disk.
After vg00 is up and running after rebooted, Can I just frestore the fbackup tapes to recover full system ?

Any one experience this situation before? Your input and comments are very appreciated. Thanks.

Please edit your post such that it's readable for others. Thank you.

Sorry for the inconvenience caused. I re-post the question as following

Attempt 1 -- To boot from normal (primary)

Unable to boot - system complains failure
SYSTEM ALERT System Name : uninitialized DATE : 10/22/2011 Time : 03/41:12 Alert Level 15 = Fatal hardware or configuration problem prevents operation Reason for alert source : 8 = I/O Source Detail : 6 = disk Source ID :0 Problem Detail : 3 functional failure

Attempt 2 -- To boot from alternate disk
Booted and go to ISL prompt and try to startup in single user hpux -is
Error display complain on the Exec Failed and couldn't find the /stand/vmunix or /vmunix

Attempt 3 -- To boot from Recovery Tape
Able to boot by select recover shell (including loadfile vgcreate vgchange lvlnboot ...)
also frecover some LVM files from fbakcup tape and all the /etc/lvmtab /etc/fstab and /etc/mnttab
successfully loaded

When try to create the vg00, following error happen

lvlnboot -v
lvlnboot : Warning Can not determine all plysical volumes on which mirrores copes of the logical volume are located:
"/etc/lvmtab" is missing Physical Volumes.
Boot definitions for Volume Group /dev/vg00
Physical Volumes beloing in Root Volume Group :
/dev/dsk/c1t0d0 (0/0/1/1.0.0) -- Boot Disk
Unable to query the I/O interface : Parameter error
lnlnboot : unable to print reliable information

pvdisplay /dev/dsk/c1t0d0

-- Physical volumes --

PV Name /dev/dsk/c1t0d0
VG Name /dev/vg00
PV Status available
Allocatable yes
VGDA 2
CUR LV 8
PE Size (MBytes) 4
Total PE 4340
Free PE 18
Allocated PE 4322
Stale PE 0
IO Timeout (seconds) default

lvdisplay -v /dev/vg00/lvol1
-- Logical volumes ---

LV Name /dev/vg00/lvol1
VG Name /dev/vg00
LV Permission read/write
LV Status available/stale
Mirror copies 1
Consistency Recovery MWC
Schedule Parallel
LV Size (MBytes) 512
Current LE 128
Allocated PE 256
Stripes 0
Stripe Size (Kbytes) 0
Bad Block off
Allocation strict/contigous
IO Timeout (Seconds) default

----- Distribution of logical volume ----

PV Name LE on PV PE on LV
/dev/dsk/c1t0d0 128 128

Logical extents --

LE PV1 PE1 Status PV2 PE2 Status 2
00000 ??? 000000 stale /dev/dsk/c1t0d0 00000 current
..

 :: Status with stale means the harddisk has problem to access

 This is one of the step require to bring vg00, without the vg00 activated successful,
 it is not possible to mount all others filesystems like /var /backup.

***************************************************************
Question is if I do not want to replace the faulty disk, I am thinking just break the mirror
and re-install the system backup (fbakcup) from tape.

Anyone has the comprehensive steps so that I can restore back all from tape? Sorry
this is the new comer for this hpux installation and recovery. Appreciate your sharing.

Thanks.

Well, since you know how to get to a recovery shell from a Ignite-Ux tape, instead of choosing second option :
Run a Recovery Shell choose the first one: Install HP-UX ...
Remove the failed disk before...
Source location choose media only installation, and Guided Installation...
While you are at it, it is a good time to think to rearrange Logical volumes size especially the 3 contiguous ones that cannot be changed after (even with OnilneJFS)

Thanks for the quick reply. You said that remove the failed disk, is this mean I need physically remove the disk or just break the mirror on the failed disk will do ?
Another question is if I choose Install HP-UX, is that mean it will install the base version and without the patches, right ? Is it possible I install base hpux and restore from frestore (to ensure all the patches are applied) ? The reason is I afraid the database and application need those patches to run and those patches are no longer exists as it been no longer support by HP. Thanks

You can patch hpux v1, patches are available on hp site if you have support.
If you don't have support, find someone who has and can download it.

QPK bundle is the thing you are looking for after restore / install.

Hope that helps.
Regards
Peasant.

So is that the only way I can recover back my OS and database ? as I afraid we don't have any support from HP as the machine is pretty old. and don't think there are anyone still using this version of OS.

Since we do not know your configuration it will be difficult to say more...
As I said remove the faulty disk to avoid confusion - or if you think ( possible ) that it is still good but not bootable, remove the sane and try...
This is a task any HP-UX admin has done a few times in his life, especially with HP-UX 10.20 and 11.00 for it is the only way to change the size of the 3 contiguous logical volumes ( /, /stand, lvol2(swap)). It was also a way to clone boxes...
If this can recomfort you:

raz01:/home/vbe $ uname -a;model;uptime
HP-UX raz01 B.11.00 U 9000/800 503746527 unlimited-user license
9000/800/L1000-36
 10:57am  up 432 days, 15:55,  3 users,  load average: 0.03, 0.03, 0.03

after I manage to boot up from recovery tape, can I just run the recover -f /dev/rmt/0m -rv (to perform full restoration to the box) rather than Install fresh HPUX OS. As I don't think I can't find the patch. Is this right approach ?

No!
Either you let the tape manage on its own (which means it will do a minimum installation (there is all the boot things to do...) then a minimum of OS is needed to continue...) in which case you boot on the tape and do nothing else till it has finished or you interact in order to customize the settings (e.g. the LVM layout). If you were happy with the settings at the time the tape was made, then let the tape do his job...
What patch were you looking for?

vbe, thanks for prompt reply.
Can I say that after I manage to boot up from tape recovery following is my tasks to perform :

1. install hpux (rather than choose recover shell) question is will it intelligent enuf
2. perform LVM layout (can restore from tape into /etc/lvmtab and /etc/mnttab )
   :: in this step I will perform lvmlnboot -v to verify, if error will use the lvreduce to break mirror.
3. Reboot the server to ensure the system able to boot from Unmirror Disk.
4. recover �f /dev/rmt/0m �rv
5. mount -a

Thanks

The tape has an option: -i if used when generating the tape... If it were not the case the tape process as you have seen can be interrupted to go in interactive mode, but this is of no use unless you want to interact with the system to modify e.g LVM layout.
If you boot from the tape and dont interfere, it will go in batch mode and every thing for you... It will take him about 3-4 hours...

That said
Your task will be after to restore every thing after using your backup software because your system (vg00 that is ) will be at the state the ignite tape was produced...
Files that may have changed are /etc/passwd ( and new accounts...), was the machine patched since? Have you created new FS since etc..

Good luck
Before If you can, try to read the /var/opt/syslog/syslog.log or OLDsyslog.log to see if it did complain about bad disk.. if not you may have had a corruption where a disk was not bootable...

Thanks vbe, after boot up using the recovery tape, I choose "Install HP Unix" and select basic installation (don't go for Advance option). Once Submit, the restoration begin and it take about 2 hours to complete. Actually I previously confuse with the "Install HP-UX" which I thought it will install the HP-UX original OS, but it fact, the recovery tape will install and recovery to the state as at the time to perform tape recovery. Thanks again.