5

using different switches). This configuration is preferable to using separate VLANs on one switch to separate them and is the configuration shown in the example cluster.

If you do not use a dedicated network as the primary network—that is, if the primary network also carries data traffic, see “If You Do Not Use a Dedicated Primary Cluster Protocol Network” and “Configuring NGX for Clustering” for configuration information.

„Secondary cluster protocol network/interface: Each node may also be connected to an (optional) secondary cluster protocol network. The interface a node uses to connect to this network is its secondary cluster protocol interface. In the example cluster, the secondary interface is eth-s4p1.

If a primary interface fails on a member, the cluster synchronization and management traffic fails over to the secondary interface. In this event, the other nodes are not affected—they continue to use their primary interfaces to communicate with the master. If a primary interface fails on the master, all the other nodes must use the secondary protocol network to communicate with the master.

If the primary and secondary cluster protocol interface fails on a node, the node is removed from the cluster. If it is the master, one of the remaining nodes becomes the new master.

These interfaces should be internal, and you should not use a secondary cluster protocol network for production traffic. If a secondary cluster protocol network fails but the primary remains functional, the cluster remains active but traffic to non-cluster devices on the secondary network might fail.

Cluster Voyager: A feature that lets you centrally manage all the nodes in a cluster as a single virtual system using one browser session.

Joining: When becoming part of a cluster, a system can copy a variety of configuration settings from another cluster node (so you don’t have to configure these settings manually). This is called joining. When a system joins a cluster, it copies the configuration settings of the join-time shared features. Joining saves you time by allowing you to configure one node and then have the other nodes copy the appropriate configuration settings when they join the cluster.

Join-time shared features: You may want to have many configuration settings be identical on each cluster node. Voyager makes this easy for you by letting you specify which features will be configured the same on all cluster nodes. The features that can be configured this way are called join-time shared features, meaning that their configurations can be shared across cluster nodes during the joining process.

Clustering Modes

IPSO clusters have three modes of operation. Nokia provides this choice so that IPSO clusters can work in any network environment. All cluster nodes must use the same mode.

Note

If you use PIM, you must use multicast mode or multicast mode with IGMP as the cluster mode. Do not use forwarding mode.

212

Nokia Network Voyager for IPSO 4.0 Reference Guide

Page 212
Image 212
Nokia IPSO 4.0 manual Clustering Modes, 212