Veritas Storage Foundation 5.0 for Oracle RAC Configuration Guide Extracts for HP Serviceguard Storage Management Suite, Second Edition, May 2008

Planning SGeRAC Installation and Configuration

About CVM and CFS in an SGeRAC Environment

Chapter 2

Coordinating CVM and CFS Configurations

After installing SGeRAC, a VCS cluster attribute (HacliUserLevel) is set to give root the

ability to run commands on remote systems by way of the cluster interconnect. CFS

takes advantage of this mechanism to enable you to perform file system operations

requiring the primary node be initiated on secondary nodes and carried out on the

primary node transparently.

If you reset this attribute, be aware of which node is the primary for certain file system

operations and perform those tasks from that node. Unlike a non-RAC environment, you

cannot run a sequence of VxVM and VxFS commands, such as resizing a volume and a

file system, on the same node unless it is both the CVM master and CFS primary node.

About Shared Disk Groups

Make sure you review the following general information when working with disk groups

and volumes. For complete details on creating and managing shared disk groups, refer to

the Veritas Volume Manager Administrator’s Guide available on the SG SMS media and

at: http://docs.hp.com/en/oshpux11iv2.html#VxVM

Viewing Information on a Disk Group

• To display information about a specific disk group, type:

# vxdg list disk_group

Checking the Connectivity Policy on a Shared Disk Group

By default, the connectivity policy for a shared disk group is set to “global.” This setting

protects against possible data corruption and causes all nodes in the cluster to detach

from the disk group when any node reports a disk failure.

The output of the vxdg list shared_disk_group command includes the following line:

detach-policy: global

• To change the connectivity policy for a disk group from “local” to “global,” type:

# vxedit set diskdetpolicy=global shared_disk_group

Determining Whether a Node is CVM Master or Slave

• On one node (nebula), determine whether the node is the master or slave:

# vxdctl -c mode

• On nebula, which is the slave, the output shows:

mode: enabled: cluster active - SLAVE

master: galaxy

• On galaxy, which is the master, the output shows:

mode: enabled: cluster active - MASTER

master:galaxy