Supported Since: 17.01
IPS uses the concept of clusters to identify, isolate and manage individual units of deployment. A cluster consists of multiple replicas of ESB instances, all with identical configurations. IPS guarantees that a cluster is maintained at the desired replication level, with new instances being spawned to take the place of any instances that get terminated.
To separate out the management and deployment aspects, configurations of a cluster are maintained as two separate types of entities:
The Cluster entity consists of basic configurations of a cluster, including
type (whether the cluster should run in
PRODUCTION mode; which will decide the
properties applied on projects deployed on the cluster
replication count (how many ESB instances should be kept running)
container image (the UltraESB-X Docker image to be used for the ESB instances spawned under the cluster)
An IPS cluster is a completely standalone management and maintenance entity from the deployment point of view. You can create, update, start, stop, refresh, monitor and delete a cluster independent of other clusters present in the same IPS deployment.
Each Cluster is associated with multiple Cluster Version entities, each of which consists of
Similar to a revision in a versioning system, a Cluster Version represents the status of a cluster in a particular deployment cycle, including the deployed project versions, configuration artifacts and port mappings (see Deploying a Cluster for more details). If you need to revert to a previous deployment configuration of the cluster, you can simply deploy a previous Cluster Version via the IPS web console.