Overview1. If a user has a cluster, in which each Exinda is licensed to 1Gbps, what is the total throughput they can optimize?
2. Can devices in a cluster be on different firmwares?
3. Are reports synced in a cluster?
4. Can a user have only one device reporting while the other optimizing?
5. How is the monitoring handled?
6. How is acceleration handled?
7. What happens when I reboot my cluster
Resolution1. The user can optimize 1Gbps of throughput in their cluster. Cluster syncs optimization. Even if one appliance is bridging (actually passing traffic through it), the other appliance makes a copy of the optimization table and replicates the flows through it. Hence limiting the total throughput to the throughput of 1 device.
2. Devices in a cluster should be on the same firmware version and their NTP time should also be synced. They also need to be the same hardware.
3. All "control" stats are synced in a cluster. That means, stats for optimization and monitoring objects. Individual System Stats and interface stats are not synced
4. Cluster monitoring cannot be conducted on a single cluster member. Both the units will monitor stats because traffic is replicated through the remote appliance as well. Both the devices optimize traffic. The one which sees local traffic impacts optimization, the one which sees remote traffic does so for reporting. In an accelerated environment, one device could be bridging the traffic and the acceleration feature might be passed over to the peer, hence load balancing.
5.The reporting/monitoring is handled by the bridging appliance.
6. The acceleration is delegated by the master. It maintains a list of all the redirects (accelerated connections) and then just cycles through it.
7. When a cluster is rebooted, the device which comes up first will become the master. The device which boots up next will be the standby. Give the devices some time to re-establish the cluster. It might take 5-10 minutes or even more in a few cases for the cluster to re-establish.