EMC

[lanke_1]

Hi,

We have an EMC symmetrix connected to two
K-boxes. One node is configured as a failover of the other.When I create LV's on VG1 of first node, the second node also has to be updated with the same configuration using vgexport, vgimport and vgchange process.

My question is when I have more VG's like VG1, VG2, VG3, etc. do I have to do vgexport and vgimport for all the Volume Groups or only VG1?

Your thoughts on this will be greatly appreciated.

Thanks in advance.
Lanke

[Christopher McCray_1]

Hello,

For failover purposes, everytime you modify the LVM information for a VG, you have to update this information on your adoptive node(s).

On node1:

# vgexport -p -s -m /tmp/vg##.map vg##

remote copy or ftp the vg##.map to the adoptive node.

on node2:

# vgexport vg##
# mkdir /dev/vg##
# mknod /dev/vg##/group c 64 0x##0000 (must be the same as before)
# vgimport -s -m /tmp/vg##.map vg##

No vgchange is done.

You should be good.

Hope this helps

Chris

[Christopher McCray_1 · Accepted solution]

In case I wasn't clear, for any and all VGs modified, this must be done, but only to those which are modified.

Sorry if this wasn't clear at first

Chris

[Uday_S_Ankolekar]

Hello,

You need to vgexport and then import only for those VG's taht have modified.
You also need to modify the service guard ascii and cntl files and compile the file (cmcheckconf and cmapplyconf)before copy it over to next node.

Good luck,

-USA...

[MANOJ SRIVASTAVA]

you would have to fo a vgexport and vgimport on the other serve for other volume groups too , I understand this is a Failover config.


Manoj Srivastava

[lanke_1]

Thank you all for your timely responses and getting my doubt cleared.
By the way are there any trouble shooting documents to which I can refer to.

Thanks
Lanke

[Christopher McCray_1]

Hello,

Go to Maintenance and support and search the technical knowledge base. Use the keyword sgux in your search. You will find quite a few docs this way, or go to

http://www.docs.hp.com/hpux/ha/index.html

For a bunch of manuals on HA

Good luck

Chris

[Rita C Workman]

You mention 'failover'...which made me think. Is your 'failover' a MC/SG failover. If it's MC/SG, then remember that after you make your lvm change...do your vgexport/vgimport...that if you have added any new vg, lvol or fs, you also need to modify your pkg script on that node and then copy the new pkg script to the other node as well.

Like I said...just in case your using MC/SG...if not then ignore this..

Rgrds,
Rita

[George Petrides_1]

The reason you are doing is, is to record any new disks in /etc/lvmtab or create device files for any new logical volumes. You could use mknod to create the device files (lvol/rvol) for a new lvol (that is if you are lazy and understand major/minor number of the device files or just export/import (don't forget the map file especially if you change the default names of the logical volumes.
Just my $0.02