root disk failure

[navin]

Hi,
I got a root disk failure in my system.The system has 2 disks.We have a ignite procedure
in environmnet.so Once the disk is fixed ,i should be able to install the system using ignite.In between i got a confusion that once the disk is replaced and formatted the ignite
process will take care of remaining activities like creating filesystems ,etc.right
Please help.

[Pete Randall]

It depends on what you backed up. The usual recommendation is to back up just your root volume group - vg00 - and use conventional backup strategies for your other data. In that scenario, you would use ignite to restore vg00, then create the device files for your other volume groups and run vgimport on them - like this:

mkdir /dev/vg01
mknod /dev/vg01/group c 64 0x010000
vgimport -s -m /tmp/vg01map /dev/vg01

On the other hand, if your Ignite backup includes the other volume groups, then everything would be restored.


Pete

[Pete Randall]

Actually, the mkdir and mknod commands may not be necessary - they may be present after the restore of vg00.


Pete

[Kyle Martin]

That all depends. If the Ignite tape was created with make_tape_recovery -x inc_entire=vg00 -x inc_entire=vg01..., then yes it will.
However, if the archive was created by doing make_tape_recovery -A, then it will only create the essential files in vg00.
The data in the data volume groups will still be there but you will need to import the vg's back in.
My recommendation is to do this procedure before restoring from ignite:

(You may have to boot into LVM Maintenance Mode for this if the boot disk won't allow you to do anything in multi user mode:
1. reboot
2. Interrupt the boot sequence within 10 seconds by pressing any key.
3. At the main menu, type boot pri isl
4. Interact with IPL? y
5. ISL> hpux -lm )

1. Get a print out of /etc/lvmtab, ioscan -fnC disk.
2. vgchange -a n /dev/vg01
3. vgexport -v -s /dev/vg01
4. Reignite the system and let it come up.
5. Once vg00 is back in place, do:
mkdir /dev/vg01
6. mknod /dev/vg01/group c 64 0x010000
7. vgimport -v -s /dev/vg01
OR
vgimport -v /dev/vg01 /dev/dsk/c#t#d# /dev/dsk/c#t#d#
8. vgchange -a y /dev/vg01
9. Mount the filesystems
10. Once the configuration is back in place correctly, then do: vgcfgbackup /dev/vg01

This should do it for you.

[Cesare Salvioni]

Kyle and Pete, seems to me that navin is talking about a ROOT disk failure, that's to say a disk in vg00 not vg01.

Navin if what i suppose is true, the host should not be working unless the second disk is itself in vg00 and hold a mirror copy for all the system LV.

In this case the host is up and running and you don't even need ignite to restore the broken disk, just change it (shutdown the host if the disks are not hot swapable) exec a vgrestore vg00 /dev/rdsk/ and a vgsync to resync the mirror copies

In the other case (os not mirrored) the host should not be working. When you apply the ignite recovery booting from the tape or from the ignite server on network, ignite process will do almost all the work for you:
create the vg00
create the LV on vg00
create the file system and recover all the files you backed up (at least the minimum for the system to start)

Anyway you should describe better how the disks were configured (whch vg they belong to) and which command (make_tape_recovery) you did. make_tape_recovery -A would backup ALL vg00

Hope this helps

[Todd McDaniel_1]

Pete/Kyle... seesm you guys are confused a bit???? he never mentioned vg01


Navin....

You say you have a root disk failure... and mentioned ignite.

However, do you have / mirrored? If so you wont have to ignite the box...

IF your system is down, your ignite tape will have saved all vg00 info and you can ignite it with little interaction...


Depending on what else you have on teh same disk as vg00, you may have to restore from other backups....