Supported Since: 17.01
A node is a working member of IPS platform. It could be a physical or a virtual machine. Instances created are actually scheduled on nodes registered in the platform. Users can add or remove nodes from IPS without restarting any of the component of the platform. This enables you to increase or decrease the computational power, memory resources and storage capacity of IPS dynamically.
IPS could have one or many nodes configured. Node Groups provide logical grouping of a set of nodes. These groups can be effectively used for advanced scheduling of clusters.
ESB instances spawned by a cluster would be localized to the set of nodes derived from the node groups assigned to it. This allows the creation of logical "regions" of deployment within the IPS platform. If a cluster is assigned multiple node groups, the "region" would be the union of nodes in all node groups.
For example, if your installation has
3 regular nodes
2 more powerful nodes
and you wish to configure 2 environments: * development: 1 powerful node, 2 regular nodes * test: all nodes
you can achieve it by defining:
dev node group containing
a test cluster (with type
TEST) belonging to the default node group
(which includes all nodes)
a development cluster (with type
DEVELOPMENT) belonging to the
dev node group
Node groups differ from zones in the respect that the latter
provides a preferential scheduling mechanism (with transition being allowed between zones when nodes become unavailable)
whereas the former is an enforcement (an instance will wait in
Pending state (refuse to start up) in case none of the
nodes derived from its node groups are available for scheduling).