about Volume Group recovery

[Platinum]

Our system had 4 HDDs, 2 VGs. 1 was vg00, the root system, contained 1 HDD. Another one was vg01, only 1 lv inside, contained 3 HDDs (3 HDDs were direct connect to system, 3 device files can be found via ioscan).
Now our root disk is crashed and unable to recovery (no system backup). We have to reinstall the root system. Will this action effect or damage the data in vg01?

[Tim Adamson_1]

Hi,

The answer is no, provided you are careful :-)

Install the OS on the replaced disk. You can always vgimport the other volume group by specifying the 3 HDSs and using the vgimport command. Refer to the vgimport man page once the system has been reinstalled.


Hope this helps.

[Rajeev Shukla]

This will not damage the data except for one thing.
After you have rebuilt the server you can connect these three disk and do a vgimport on all 3 disks you'll get vg01 and the LV also.

Now if you had some other name for that LV than lvol1 that name will change to lvol1 and you'll have to use that now . Rest every thing will be intact and no data on this vg01 and lv will be lost.

Rajeev

[Gerrit Beyken]

Hi,

you can preserve your LV-name by using the mapfile-option of the "vgimport".
Just create an mapfile with "vi" like this:
1 lvol_test
2 lvol_test2
3 lvol_test3

Then do an vgimport -m .

regards

Gerrit

[Helen French]

The first thing I would suggest is to start taking backups of your system once everything is back and ready. You don't need to worry about data loss if you do that. Download and install Ignite-UX (for root disk backup) and use fbackup for data backup (if you don't have a special backup software like OB).

To answer your question, No, a root disk crash will not erase data on VG01. You need to create /dev/vg01 directory, create group file with mknod command and then vgimport the old VG. If you want to find out which disks are part of VGs do a vgscan -a. Then you need to activate VG with vgchange -a y and check all logical volume status. Then you can mount it after creating mount points (mkdir).

If you had a recovery tape from Ignite-UX, all you had to do is replace failed disk and boot from recovery tape!

[blal · Accepted solution]

Hi

Better you disconnect all the disks in the VG01
volume group before installing new OS.You replace the failed HDD and better to set the same SCSI ID as same as the failed one.

1)Insatll OS on new disk.
2)Power off server , connect all the VG01 disks previously removed.
3)Boot server , using ioscan verify that all the disks are getting detected and device file names are created.
4)mkdir /dev/vg01
5)mknod /dev/vg01/group c 64 0x010000
6)vgimport /dev/vg01 /dev/dsk/disk1devicefile /dev/ds/disk2devicefile /dev/dsk/disk3devicefile
7)see /dev/vg01/lvol1 device file will get created if import has gonethrough fine.
8)vgchange -a y /dev/vg00
9)mount /dev/vg01/lvol1 /mount-point.
10)If required update the /etc/fstab.


But its recommented to take the root filesystem backup atleast once in a month.You can down load Ignite-Unix from HP site and create recovery tapes which becomes bootable tape .

Or if you have support+ CD with you there is one ODE copy utility trough which also you can take bootable tape image of ur system disk.

regds
baiju.