In order to attain high availability and load balancing, it is possible to configure multiple midPoint nodes working with the same repository.
In the following figure, there are two such nodes (A and B).
These nodes then share the load in the following ways:
- MidPoint tasks (e.g. live synchronization, reconciliation, import, workflow approvals, etc) can run on either node.
- Interactive users can connect to either node.
There can be two or more nodes in such a cluster. (Although the number is not limited, some practical limits probably exist; they have to be determined experimentally.)
- For task load balancing, the Quartz scheduler is used (see ).
- For load balancing for interactive users, standard tomcat solution is recommended.
- For fail over without load balancing for interactive users network-level setup with Virtual IP may be used.
For practical information on forming a cluster, see the administration guide.