Adding a Disk to an existing Vg that belongs to ServiceGuard

[Juan Manuel Naranjo A.]

Hello,

We are going to add a disk and create new Logical Volumes in an existing Volume Group that is shared between two(2) nodes on a cluster. Below is a sketch of what we would do (Please tell me if I’m missing any steps). The Logical volumes are going to be Raw Devices to be used by ORACLE.



1- Shut down ORACLE RAC on both nodes
2- Node B: # cmhaltpkg < package >.
3- Node B : # cmhaltnode
4- Node A: # vgchange -a e /dev/vgxx on node A
5- Node A: Do our work ( Add new disk, create new Logical Volumes)
6- Node A : # vgexport -p -s -m /tmp/vgxx.map /dev/vgxx
8- Node A: # vgchange -a n /dev/vgxx
9- Node A: # vgchange -c y /dev/vgxx
7- Copy /tmp/vgxx.map file to node B
8- Node B : # vgexport -m /tmp/vgxx.old_mymap /dev/vgxx (This will delete Vg on Node B )
9- Node B: : # mkdir /dev/vgxx
10- Node B : # mknod /dev/vgxx/group c 64 0x0xx000
11- Node B: # vgimport -v -m /tmp/vgxx.mapfile vgxx
12- Startup Serviceguard Nodes and Services on both nodes
13- Startup ORACLE instances.


Note: I have searched for Document Ids UXSGLVKBAN00000002 and UXSGLVKBAN00000008 and haven’t found them.

[HGN]

Hi

All the steps look good , we did somethign similar to this recently on 2 separate clusters.

Have to make sure you can see the new disks on both the servers.

Rgds

HGN

[Redhat · Accepted solution]

As you need downtime halt the cluster .

1- Shut down ORACLE RAC on both nodes
2- Node A: # cmhaltcl -f
3- Node A : # vgchange -c n /dev/vgXX
4- Node A: # vgchange -a y /dev/vgxx on node A
5- Node A: Do your work ( Add new disk, create new Logical Volumes)
6- Node A : # vgexport -p -s -m /tmp/vgxx.map /dev/vgxx
8- Node A: # vgchange -a n /dev/vgxx

7- Copy /tmp/vgxx.map file to node B
8- Node B : # vgexport -m /tmp/vgxx.old_mymap /dev/vgxx (This will delete Vg on Node B )
9- Node B: : # mkdir /dev/vgxx
10- Node B : # mknod /dev/vgxx/group c 64 0x0xx000
11- Node B: # vgimport -v -m /tmp/vgxx.mapfile vgxx
12- #Node A: vgchange -S y -c y /dev/vgXX
vgchange -a s /dev/vgXX (as well as in Node B)
13- cmruncl -v
12- Startup Serviceguard Nodes and Services on both nodes
13- Startup ORACLE instances.

NOte:as you are having RAC the vg should be shared by both the node

Hope it will help ...

[Redhat]

Before starting the oracle instances make sure to chown oracle:dba /dev/vgXX/rlvol* on both the node

[Juan Manuel Naranjo A.]

Solution was give pretty fast