What is the accepted level of network latency between different nodes in a ClearPass cluster to synchronize without errors over a WAN connection?
Environment : Cluster running ClearPass 6.0 to 6.4.x versions in a WAN network
Aruba has tested up to 200 ms and seen cluster sync to be fairly stable with the following caveats:
- As the network latency approaches this limit or exceeds it, it could be noticed that the subscribers lag behind, especially during periods of activity and provisioning/configuration changes on the publisher. The subscribers will attempt to catch up once activity dies down.
- At latencies much higher than this number, the subscribers may never catch up completely and will get dropped from the cluster eventually.