Volume Group

[Karthik Narayanan_4]

Hi,
I have a 2 node cluster running MC/SG & HP11i.

I have problems with export/import of VGs.

I created a volume group /dev/sharedvg. It was activated on Node1. Now using LVM commands, I distributed to node2. Now I deactivated on Node1. Now when I try to activate on Node2 , i got the error:
vgchange: Activation mode requested for the volume group "/dev/sharedvg" conflicts with configured mode.

Now, again i come to Node1 and try to activate. I get the same error as above.

Please help.

Thanks & Regards,
Karthik Narayanan

[Sridhar Bhaskarla]

Hi Karthik,

If it is already part of serviceguard, the preferred mode of activation is 'vgchange -a e' so that you don't change the VG configuration.

Now if you just created the volume group, I don't see any reason why it would give this error.

See if the following helps you,

On the primary node

#vgchange -a n /dev/sharedvg
#vgchange -c n /dev/sharedvg

Then on the secondary noe

#vgchange -a y /dev/sharedvg

-Sri

[Jeff Schussele]

Hi Karthik,

How are you activating it?

Use

vgchange -a e /dev/vg_name

e in this case means exclusive.
The other thing to be aware of is the cluster bit. If this VG will be used by MC/SG you'll need to do the following

vgchange -c y /dev/vg_name

To make the VG cluster-aware.

HTH,
Jeff

[Tim D Fulford]

I've seen this message in two states

1 - SG up and running and I typed
# vgchange -a y sharedvg
This is wrong, it should be
# vgchange -a e sharedvg

2 - SG was down (cluster, node, package.. EVERYTHING). I typed
# vgchange -a y sharedvg
As SG is disabled, cmcld is not running hence, I should have typed
# vgchange -a y sharedvg

You probably needed to do 1) but 2) is useful to know if you are in trouble & want to start the cluster services "manually" WITHOUT ServiceGuard.

Tim

[Tim D Fulford]

OOOPS 2 should read
# vgchange -a e sharedvg
wrong...
# vgchange -a y sharedvg

sorry for the confusion

[Karthik Narayanan_4]

I do not have SG running. Everything is down.

[Tim D Fulford]

I want a re-call answer option!!!.... Here is what I meant to write

I've seen this message in two states

1 - SG up and running and I typed
# vgchange -a y sharedvg
This is wrong, it should be
# vgchange -a e sharedvg

2 - SG was DOWN (cluster, node & package.. EVERYTHING). I typed
# vgchange -a e sharedvg
As SG is disabled, cmcld is not able to communicate with other nodes in cluster hence, I should have typed
# vgchange -a y sharedvg

You probably needed to do 1) but 2) is useful to know if you are in trouble & want to start the cluster services "manually" WITHOUT ServiceGuard.

Tim

[Karthik Narayanan_4]

Now, I blew off everything on the disk.
vgexport /dev/sharedvg

Cluster is DOWN.
Now on node1, I create physical vol
pvcreate -f /dev/rdsk/c4t9d0

Now on node1, I create VG, LV, Filesystems using SAM.
All is fine, everything is mounted.
Now i deactivate on node1, by
vgchange -a n /dev/sharedvgNEW

Gets deactivated.
Now, I try to activate again on node1 by
vgchange -a y /dev/sharedvgNEW.
I get error message as put on the previous post by me.

[Tim D Fulford]

If SG is down (EVERYTHING)

vgchange -a y sharedvg

REMEMBER, you will be doing this outside of SG, so if you fix the problem you will need to stop package processes, unmount filesystems & de-activate sharedvg (possibly remove shared memory segments if left lying 'round !!!)

Tim

[Tim D Fulford]

If SG is down (EVERYTHING)

vgchange -a y sharedvg

REMEMBER, you will be doing this outside of SG, so if you fix the problem you will need to stop package processes, unmount filesystems & de-activate sharedvg (possibly remove shared memory segments if left lying 'round !!!) Before starting SG again.

Tim