vgexport before make_tape_recovery

[mura-kei7]

Would you please tell me whether vgexport must execute before online make_tape_recovery?

Why are dis-activation(vgchange -a n /dev/vg00) and vgexport required to backup?

[James R. Ferguson]

Hi:

There is no need to do any 'vgexport' before a 'make_tape_recovery'.  If you wish to exclude and/or include volume groups you can use the '-x' options.  See the manpages for 'make_tape_recovery'.

Regards!

...JRF...

[mura-kei7]

Thank you for your advice,

>There is no need to do any 'vgexport' before a 'make_tape_recovery'.

In case of online backup or single user mode, is the vgexport unnecessary?

I have the backup manual using make_tape_recovery written by someone,

do vgexport before make_tape_recovery.....

Regards,

[James R. Ferguson]

---

@mura-kei7 wrote:

In case of online backup or single user mode, is the vgexport unnecessary?

I have the backup manual using make_tape_recovery written by someone,

do vgexport before make_tape_recovery.....

---

A backup (including 'make_tape_recover' which usees 'pax' to perform its file backups) targets a directory or directories of a filesystem.  in that case, the filesystem must be mounted (unless you are using 'dd' to copy the physical disk).  Mounted filesystems, of course, mean active volume groups.  The only thing that matters when you take a backup is that the objects being backed-up are static (un-changing).

Re-read my first response: I have no idea why someone specified a 'vgexport' (of what?) before a 'make_tape_recovery'. 

When you run 'make_tape_recovery' a "preview" ('-p') 'vgexport' is actually performed for every volume group to create a current mapfile in '/etc/lvmconf'.  A 'vgcfgbackup' is also performed.  Together, these actions provide you the basis for a 'vgimport' or a 'vgcfgrestore' should either be needed.  Perhaps this is what the notes you were left meant to say.

Regards!

...JRF...

[mura-kei7]

Hi,

In the manual(someone wrote),

1.vgexport -v -p -m /vg00.map /dev/vg00

2.make_tape_recovery -p -A -a /dev/rmt/0mn

3.vi /var/opt/ignite/recovery/config.local

   vgreduce, insf, pvcreate, mkboot, ,,,

4. vi system_cfg and add "init _hp_ignote_sw_impact=1"

5. vi os_arch_postl and change "merge_file" to "save_file"

6. make_tape_recovery -r -v -a /dev/rmt/0mn

If doing just 6.make_tape_recovery, 1-5 steps are unnecessary?

[James R. Ferguson]

Hi:

Step-1 is done by 'make_tape_recovery' as I indicated so this is a waste of effort.

Steps 2-6 are a sequence to allow you to create customizations to the recovery archive to be created.  Only you and decide what/why you might want to do this.  Commonly, a 'make_tape_recovery' is made simply for vg00 with something like:

# /opt/ignite/bin/make_tape_recovery -x inc_entire=vg00 -I -v -a /dev/rmt/0mn

Regards!

...JRF...

[donna hofmeister]

a read of this might be helpful:

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02023840/c02023840.pdf

vgexport in *preview mode* (-p) is very different from doing a real vgexport.  having an up-to-date map file of ALL your volume groups is useful -- although i think doing vgcfgbackup's might be better advised.

the 1st time make_tape_recovery is run, the "-p" says to run it in preview mode -- which creates the files which your instructions then say to edit.  like james says -- you (and your co-workers) need to know why this is being done.  it seems to me that these instructions may be old and/or were written to deal with a specific situation (that may no longer be relevant).

i fully agree with james' recommendation for how to make an ignite tape.

and you are checking afterwards to see if the tape is useable, yes? :-)