About policy settings for endpoints

On the Controller Settings subtab, you can create an ordered list of controllers available to endpoints. A controller’s position in the list determines when endpoints attempt to connect to it; endpoints attempt the first controller, and then move down the list until they connect. If you are using controller clusters, you can enable auto-update to have the list here updated with new cluster members as they join. If you want to enable endpoints to connect randomly to available controllers, you can enable that option, overriding the list order.

The Insert At setting enables you to specify a newly added controller’s position in the list of controllers. Select start, end, or a number from the drop-down list. The default value is End. Service instances connect to controllers according to the controller’s position in the list, starting with 1. If the service is unable to connect to 1 in the list, the system moves on to the next controller in the list. For example, if the system is unable to connect to controller 1, then controller 2 is attempted. If controller 2 is successful, no further controllers in the list are attempted.

The Hostname/Port setting specifies the fully qualified hostname or IP address and port for the controller that the client connects to. You can specify more than one controller. The default port value is 7870.

Enable the Use Random Ordering of Controllers when Connecting enable endpoints to randomly connect to any available controller, disregarding the priority list. The setting is disabled by default.

In cluster environments, remove the default entry for this setting. When the default entry is removed, the system automatically uses the correct fully qualified domain name for the controller that sends the policy.

On the Desktop Settings subtab, you can specify endpoint settings like the data store size, log size, Windows-only settings, and enabling visibility of the Riverbed icon on the endpoint. Endpoints must have a minimum of two gigabytes free storage space for client software installations to succeed.

When installing client software on endpoints, Client Accelerator allocates endpoint storage at the configured amount to a single, empty file which is used for the endpoint’s data store. The data store is used by the optimization service and it stores information shared with peer SteelHeads. On endpoints that do not have enough free storage to meet the amount specified in the policy, Client Accelerator automatically adjusts the data store size so that the client software installation succeeds (providing the endpoint has 2 GB or more free storage space). This is called fallback data store resizing. After installation, if enough endpoint storage space is made available to accommodate the original data store size specified in the policy, Client Accelerator automatically performs a one-time upgrade of the data store to the policy setting, which clears the data store contents. After upgrade, the Client Accelerator resumes collection of optimization data.

Fallback data store resizing is disabled by default. You can enable the feature by using the policy id <X> endpoint autodg-segstore enable command. For details, go to Knowledge Base article S36068.