on the target!!!!

[hpuxhelp]

The root vg is mirrorred. After breaking the mirror and remove the secondary path from the vg to ensure the good copy of the O/S is still maintained, I should be able to boot to either of the disk? Is it correct?

[James R. Ferguson]

Hi:

Yes, assuming that you setup the mirror correctly in the first place.

Before you break the mirror and boot, check with 'lvlnboot -v'. The output should look like:

Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c2t6d0 (10/0.6.0) -- Boot Disk
/dev/dsk/c2t5d0 (10/0.5.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c2t6d0
/dev/dsk/c2t5d0
Root: lvol3 on: /dev/dsk/c2t6d0
/dev/dsk/c2t5d0
Swap: lvol2 on: /dev/dsk/c2t6d0
/dev/dsk/c2t5d0
Dump: lvol2 on: /dev/dsk/c2t6d0, 0

Regards!

...JRF...

[Jeff Schussele]

Hi,

Yes you should.
But I'm not sure what you mean to accomplish here.
Once you break the mirror the "alternate" disk will no longer be "equal" to the "primary" disk. i.e. any changes made to primary will NOT be reflected on the alternate. So be careful just which disk is which & boot the one you want.

Rgds,
Jeff

[hpuxhelp]

the problems that I ran into was that once I did break the mirror and vgreduce the secondary disk, I can't boot it to secondary device? Why...
Breaking the mirror and reduce the disk, keep the intergity of the good copy...it should work if I decided to boot to boota or boot b disk?
** lvlnboot display the both secondary and primary boot...
** after lvreduce command , the only disk left was the primary one...make sense...

*** why can't I boot to the secondary disk after breaking the mirror.. and remove this secondary disk out of the v

[Tim D Fulford]

I hate to contradict the above, but If you have vgreduced the mirror OUT of VG00, that means that the LVM info in the header of the disk will not be correct (how many disks are in the group vgxx??)

I have to say I've never tried this, I'm just guessing from previous experience. It is just that the VGRA & VGDA (kept in the header & footer of the disks) will probably be out of date & probably incosnistent.

If you want to check that the VG info is correct/consistent on the REDUCED disk try vgimporting it in so

# mkdir /dev/vg00old
# mknod /dev/vg00old/group c 64 0x050000
# vgimport /dev/vg00old /dev/dsk/cxtydz
If this volume group imports, good, you can also check it a bit deper by mounting the lvols up.

# mkdir /oldroot
# mount /dev/vg00old/lvol1 /oldroot
# mount /dev/vg00old/lvol3 /oldroot/stand
etc...

The next test you can do is check the boof/lif info
# lifls -l /dev/dsk/cxtydz
# lifcp /dev/dsk/cxtydz:AUTO -

IF all the above is OK I would say that it should be able to boot...

also note all the above can be done whilst the system is ON-LINE just to make sure you are sane

[hpuxhelp]

very funny......!!!!:)
but you are right on the fact that I am not insane eheh :)
never mind...I don't know what I was thinking....
I think I am now understand what was my the problem...
If the pv was remove from the volume group, it is still have its data, then in the vgxx has only one disk. When you boot you boot on the disk that still in the volume group. Once that disk is successfully be upgrade or whatever, you can add the secondary disk back to create the link.... eeheheh I was out somewhere when I asked this question...
thanks

[Steven E. Protter · Accepted solution]

If you have to do it over again...

Here is a procedure that results in the outcome shown in post 2 above.

Put in real numbers instead of the #'s

pvcreate -B /dev/rdsk/c#t#d#

mkboot /dev/rdsk/c#t#d#

mkboot -a "hpux -lq (;0)/stand/vmunix" /dev/rdsk/c#t#d#

vgextend /dev/vg00 /dev/dsk/c#t#d#

lmextend -m 1 /dev/vg00/lvol1 /dev/dsk/c#t#d#

Repeat for any other volume groups.

For root
lvnlboot -r /dev/vg00/lvol3

Swap
lvnlboot -s /dev/vg00/lvol2

dump
lvnlboot -s /dev/vg00/lvol2

For 10.20 and 11.00 only
lvnlboot -b /dev/vg00/lvol1

Hope this is usefull, I pulled it out of a support call. I'm sorry I don't have a document id.

Steve

[James R. Ferguson]

Hi:

Have a look at Technical Knowledge Base document #rcfaxlvm007 - "Booting From a Split Mirror".

Regards!

...JRF...

[hpuxhelp]

Please help me to clarify these questions:
1. if the secondary boot device was removed from the vgxx, the vgxx is now has no mirror and only 1 disk and that is the primary boot disk, if I can change the setboot -p , will the system boot to the secondary boot device? If this boot then what if
2. If the upgrade of the patches somewhat cause the corruption on the secondary boot, and can't longer boot, how do I redo the mirror using the old copy from the original primary disk?