Optional: Increasing fault tolerance

By default, the multi-node HA-ready production profile allows for one server node going down without causing a deployment failure. You can increase the tolerance to server node failure using the fault_tolerance parameter in the cluster_config.json file. The parameter modifies the replication factor of in-cluster storage components such as Ceph and Longhorn.

Hardware requirements

To increase fault tolerance beyond 1, make sure your environment meets the following requirements:

• Your cluster consists of a minimum of 2x+1 server nodes, where x is the required server node fault tolerance;
• Each server node has raw device configured.

How to increase fault tolerance

To increase fault tolerance beyond 1, take the following steps:

1. Set fault_tolerance to the required value in the cluster_config.json file. If you set it before starting the installation or upgrade operation, you do not need to take any additional steps.
2. Run the uipathctl.sh installer to modify the in-cluster Ceph objectstore replication factor. Wait until the operation completes successfully.
3. Run the install-uipath.sh installer to modify the Longhorn volumes replication factor. Wait until the operation completes successfully.