The service cluster and the advanced networking functionality are the best additions to this version of Zen. In the next section we will explain the concepts and functionalities of this service.
Zen Load balancer cluster architecture.
- eth1: Cluster Real IP, You configure this ip on web GUI, and this configuration IP will not be replicated for backup node. The cluster service on node2 checks the Real IP on node1, and the cluster service on node1 checks the Real IP on node2. Also you can use a cross over cable for configure this siervice, then you shouldn’t configure farms on these interfaces
- eth1:cl : Cluster Virtual IP, only active on active node, You can configure farms over this IP if you don’t use a cross over cable
- UCARP: Checks the real ip of cluster nodes. UCARP runs a virtual ip on the active node.
- Zen inotify: Service replication, this is only running on the active cluster node. This service replicates the entire configuration for the BACKUP node
- Communication takes place between nodes over RSA, and replication configuration over inotify
- The Real IP will not be copied to the backup node. Interface configurations, route tables and configuration data for farms will be copied to backup node with rsync (RSA). Any time you modify a file Zen inotify runs and all configuration files will be replicated to the backup node. In other words only set your farm configuration details on the MASTER node, Zen will replicate your settings to the BACKUP node.
- You can use the eth0 (weggui network interface) for configure virtual interfaces, like eth0:1 and farms.
- If on your real environment you only have a network (eth0) you can configure webgui, cluster service and farms over the same interface following the example with no problem.
The cluster service requires two Zen Load Balancer appliances. One will run a real ip and the other a virtual ip on the same subnet. You can configure this service under Settings >> Cluster. When you configure and run the cluster service the interfaces will be locked so make sure that you are happy with the network configuration before you enable the cluster service.
RIP hostname1: 172.16.1.10 over eth1
RIP hostname2: 172.16.1.11 over eth1
VIP cluster: 172.16.1.100 over eth1:cl