Operating System - HP-UX
1830340 Members
2559 Online
110001 Solutions
New Discussion

Service guard failover Question...

 
SOLVED
Go to solution
jmckinzie
Super Advisor

Service guard failover Question...

I have migrated from one SAN to a new SAN....

Now, I need to verify package failover works....however, I have an issue...

on the primary host I have packages that run on controller C5txdx on the next server, the txdx is the same however the controller is different. How will service guard know to mount the vg on the new controller in the event of failover?

-TIA
6 REPLIES 6
Luk Vandenbussche
Honored Contributor
Solution

Re: Service guard failover Question...

You have to update the VG info on the second node.

vgexport vgxx
mkdir /dev/vgxx
mknod /dev/vgxx c 64 0x0x0000
vgimport vgxx /dev/dsk/cxtydz
vgchange -a r vgxx
vgcfgbackup vgxx
vgchange -a n vgxx

This can be done while the package is running on the first node
Geoff Wild
Honored Contributor

Re: Service guard failover Question...

As long as you updated your cluster ascii file and cmapplyconf, then it should work.

I'm assuming you also did a vgexport/vgimport on the non active node after the SAN migration.

Rgds...Geoff
Proverbs 3:5,6 Trust in the Lord with all your heart and lean not on your own understanding; in all your ways acknowledge him, and he will make all your paths straight.
jmckinzie
Super Advisor

Re: Service guard failover Question...

Luk,

I assume the vgexport vgxx is exporting the volume group?

Is this correct?
HGN
Honored Contributor

Re: Service guard failover Question...

Hi

You can export the vg in preview mode that woudl create a map file, import if after you have exported the already existing vg in the same name.
once this process is complete the failover will work fine,we have done this many times in our site
Make sure the disks are seen on both the servers.

RGds

HGN
Mridul Shrivastava
Honored Contributor

Re: Service guard failover Question...

You need to export the vg in preview mode using the following command:

vgexport -p -s -m /tmp/vgname.map /dev/

This will create a file under /tmp that you have to ftp to other nodes of the cluster and then run vgimport... This will populate the lvmtab with the vg information and it will take care of the disks naming as well since we have specified -s
Time has a wonderful way of weeding out the trivial
jmckinzie
Super Advisor

Re: Service guard failover Question...

I have completed this issue based on feedback provided above.