Bootable disks failed, moving the data to other server

[csreenivas]

We have both the bootable disks failed on HP UX 11.0(server 1) and we would like to move the data to another HP UX 11.0(server 2) which is running but no importance.
Leaving bootable disks we want to destroy data on server2 and restore the data(server 1) from Tape. We have full backup for both the servers Can you please send me the proceedure.

We have to recreate 2 volume groups and logical volumes on server 2.

[Ganesh Saraf]

forstly Do you have;
Following thing which are important before starting installation;
1) Ignite backup for server 1
2) vg configuration backups of both VG's.


if you dont have this you need to frshly install OS. And recreate both VG's with own specification of size for respective LV's in vG's.
if you have ignite just restore the ignite bckup to server 2. Aftwer this restore vg configuration backup to disks you to restore (dreate vG's) using vgcfgrestore ( assuminbg that both VG's which you want to restore are non root VG's)

feel free to post problem if any..

[csreenivas]

We dont have ignite backup and also dont have OS (11.0) cds to install. We dont like to disturbing the OS on server 2, we want to erase data on server 2 and copy the data from server1 using tape(ordinary backup).

[Johnson Punniyalingam · Accepted solution]

Hi Sreenivas,

data means--> are you refering to "vg01" apart from root file system..?

can you post-output for bdf,vgdisplay -v

if its from apart from root file system, steps are simple

Server2,
========
# umount /filesystem under vg01
# lvremove /logical_volume under vg01
# lvcreate -L logical Volume under vg01
# mount /file_systems under vg01 --> as per the file_system (same as the Server1)

Than start to restore the data from tape, on the new-created logical volumes under server2.

Thanks,
Johnson

[Ganesh Saraf]

Hi Sreenivas,
As i can understand from your last post; you want to earse all the data on VG other than root VG.
Considering this scenario;
We will assume following things for non root VG;
OLD VG details:
VG Name: vg01
Lvname in VG01: /dev/vg01/lvol1 monted on /gan and /dev/vg01/lvol2 mounted on /shree
Disk included in VG01: /dev/dsk/c0t1d0 (assuming 18 GB disk)

NEW VG Details:
VG Name: vgnew
LV Name: lvnew1
2nd LV Name: lvnew2
disk: /dev/dsk/c0t1d0 (same as above)

Now The STEPS:
-# umount /gan
-# umount /shree
-# vgchange -a n vg01
-# pvcreate -f /dev/rdsk/c0t1d0 (ALL DATA WILL BE LOST)
-# vgcreate newvg /dev/dsk/c0t1d0
-# lvcreate -L 9000 -n lvnew1 vgnew
-# lvcreate -L 9000 -n lvnew2 vgnew
-# newfs -F vxfs /dev/vgnew/lvnew1
-# newfs -F vxfs /dev/vgnew/lvnew2

After this you can proceed for data restore from tape.

Reply back if you face any problem

cheers..

[csreenivas]

Thanks for the information. we have created logical volumes on vg01, vg02. we have succesfully restored the data on vg02. But while restoring to vg01 using frecover we are getting the below error. Can you help me in this?

vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
frecover(1048): I/O error in write of file /dev/vg01/data_0
-rw-r--r-- sybase sybase /dev/vg01/data_0
frecover(1073): frecover of data_0 failed

[OldSchool]

vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
frecover(1048): I/O error in write of file /dev/vg01/data_0
-rw-r--r-- sybase sybase /dev/vg01/data_0
frecover(1073): frecover of data_0 failed

it appears that you "recovered" the data to a non-existent device file, so it went into the /dev directory and filled the root drive.
either you missed something creating the filesystems (can't tell since you didn't provide any details), or you goofed something up creating the backup (again no details).

[Johnson Punniyalingam]

Hi Sreenivas,

>>>vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
frecover(1048): I/O error in write of file /dev/vg01/data_0
-rw-r--r-- sybase sybase /dev/vg01/data_0
frecover(1073): frecover of data_0 failed<<

Check step1 :- bdf --> look for any file system full reached 100%
Check step2 :- frecover command you need point the

Thx,
Johnson

[csreenivas]

Thank You. We have restored the data to vg02 mount point and the DBA team has started database. At last 50% of the problem has been solved after trying for many options. Learned many new things and realised the importance of ignite backup.

[OldSchool]

">>>vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
frecover(1048): I/O error in write of file /dev/vg01/data_0"

"Check step1 :- bdf --> look for any file system full reached 100%"

well, it told you "/dev/root" was full....that's another name for "/". IIRC, tt's usually created when you when you drop back to single user mode from the normal run level, or if you transition from single user mode to the std run-level without a re-boot...I don't recall which exactly.

[csreenivas]

Below are some of the logical volumes in server 1 and we have created the same logical volumes in server 2 but when we try to restore the data to the corresponding lvs we are getting errors.

LV Name /dev/vg01/lv22
LV Status available/syncd
LV Size (Mbytes) 200

LV Name /dev/vg01/lv23
LV Status available/syncd
LV Size (Mbytes) 200


vxfs: mesg 001: vx_nospace
- /dev/root file system full (1 block extent)
frecover(1048): I/O error in write of file /dev/vg01/data_0
-rw-r--r-- sybase sybase /dev/vg01/data_0
frecover(1073): frecover of data_0 failed

our root file system is not full(58%).

I have used the below command for restore, can you please send me the right command to restore data for the above lvs.
And we have created vg01 on /dev/, as per the server 1, is it correct?

#frecover -xvf /dev/rmt/0m -i /dev/vg01

[Johnson Punniyalingam]

Hi Sreenivas.

Please follow has below

>>vxfs: mesg 001: vx_nospace
- /dev/root file system full (1 block extent)
frecover(1048): I/O error in write of file /dev/vg01/data_0
-rw-r--r-- sybase sybase /dev/vg01/data_0
frecover(1073): frecover of data_0 failed<<

step1 :-
=====
# mv /etc/mnttab /etc/mnttab.old
# mount -a
usually it should look like " / " this happen becoz when we startup the server / recover server from "Single User mode"

step2 :-

>>>>frecover(1048): I/O error in write of file /dev/vg01/data_0<<
<<

above error is due to point to logical volume ,

>>frecover(1048): I/O error in write of file /dev/vg01/data_0<<

# frecover -xvf /dev/rmt/0m -i /new_created_mount_piont

[csreenivas]

Thank You Johnson, Can you please check the restore command,

#frecover -xvf /dev/rmt/0m -i /dev/vg01/lv90

Data was not copied inot the corresponding logical volume.

/dev/vg01> frecover -xvf /dev/rmt/0m -i /dev/vg01/lv90
/dev/vg01> bdf lv90
Filesystem kbytes used avail %used Mounted on
lv90 409600 1205 382871 0%

[Ganesh Saraf]

Hello Sreenibas,
I think you have made some mistake while taking backup of that LV. where it was mounted on old server?? If your LV name is /dev/vg01/lvol1 and if it is mounted on /shree

your command to take the backup of this LV whould be;
-# fbackup -0v -f /dev/rmt/0m -i /shree
not like follwoing command;
-# fbackup -0v -f /dev/rmt/0m -i /dev/vg01/lvol1