Extending the Veritas volume in service guard cluster

[Senthil_N]

Hi All,

The requirement is to extend the veritas volume that is shared between two nodes in Active / Passive SG clsuter. The disks are CDSDISK.

File system / volume name:

#bdf /test

/dev/vx/dsk/abc00/test
3523 3303 2056 94% /test

DG and disks details part of both nodes:

Node1:

# vxdisk list | grep -i abc00
disk239 auto:cdsdisk disk239 abc00 online
disk240 auto:cdsdisk disk240 abc00 online
disk241 auto:cdsdisk disk241 abc00 online
disk244 auto:cdsdisk disk244 abc00 online
disk243 auto:cdsdisk disk243 abc00 online
disk245 auto:cdsdisk disk245 abc00 online

Node2:

# vxdisk -o alldgs list | grep -i abc00
disk120 auto:cdsdisk - (abc00) online
disk119 auto:cdsdisk - (abc00) online
disk118 auto:cdsdisk - (abc00) online
disk129 auto:cdsdisk - (abc00) online
disk123 auto:cdsdisk - (abc00) online
disk124 auto:cdsdisk - (abc00) online

Right now the volume "test" is part of service guard package and is mounted on Node1. 

Now It is fine if we assing the LUN to both nodes. And add the disk to the DG "abc00" and extend the volume "test" on Node1 only. Will it be replicated on Node2 automatically or do we need to run any command to get this change replicated on Node2 and Will it work fine if the package is failed over to Node2?

Please let me know if there is any change in my understanding.

[Patrick Wallek]

I'm thinking about this in LVM terms, but the same should apply to VxVM.

If you are just increasing the size of a volume (plex?) then no further action should be necessary on the secondary node.  That should come up fine.

If you need to add a LUN to the VG (disk group) before increasing the size then you will need to vgexport/vgimport (deport / import the disk group) on the secondary node so that the VG / disk group know that there is an additional disk configured.

[Stephen Doud]

Since the DG is not  identifed as a 'shared' DG, it is deported when the package is halted, and imported (and registered) when it is imported through package startup, so the VxVM metadata on the disk group is read into VxVM and the new volume size will be known.