3.6.0 Matrix Server Command Reference (5697-7083, December 2007)

Chapter 3: mx Commands 111

add—Add a new Virtual CIFS Server

mx vfs add [--policy autofailback|nofailback] [--application

<application name>] <ip> <name> ALL_SERVERS <primary network

interface>| <network interface> ...

<ip> <name> asks for the IP address and hostname for the Virtual CIFS Server.

The hostname can include up to 32 characters.

If you specify ALL_SERVERS, the Virtual CIFS Server will be defined on all of

the servers. You will also need to specify the primary interface for the Virtual

CIFS Server, as it determines the order of the interfaces for failover.

If the Virtual CIFS Server should be defined only on certain nodes, you can

specify the appropriate network interfaces on those nodes. The order in

which you specify the network interfaces determines the ordering for failover.

The first network interface is the primary interface; network traffic for the

Virtual CIFS Server goes to the node providing this interface. The remaining

network interfaces are backups. Each network interface must be located on a

different node.

NOTE: If a Virtual CIFS Server is configured with ALL_SERVERS, any new

nodes added to the matrix are also added automatically to the Virtual

CIFS Server. The PolyServe Management Console does not specify

whether a Virtual CIFS Server has been configured with

ALL_SERVERS. In addition, you cannot use the Properties or Rehost

options on the PolyServe Management Console to remove a node from

a Virtual CIFS Server configured with

ALL_SERVERS. Instead, you will

need to use the mx vfs update command to reconfigure the Virtual

CIFS Server on the appropriate nodes.

The arguments are:

[--policy autofailback|nofailback]

This argument determines the failback action that the Virtual CIFS Server will

take following a failover to a backup node.

• AUTOFAILBACK. This policy is intended to return the Virtual CIFS Server

to its original configuration, or as close to it as possible. After the Virtual

CIFS Server fails over to a backup node, the ClusterPulse process watches

the health of the nodes that are higher in the list of servers configured for

that Virtual CIFS Server. When the health of one of these nodes is equal to