OverviewIn Cluster and high availability network architecture on 10k devices, It is observed that performance degradation occurs across the network when traffic exceeds the Exinda applianceÕs Packets Per Second (PPS) limit i-e 687,000 PPS (this pps limit is only for clustered environment). This will result in significantly reduced Internet browsing speeds.
In order to avoid the performance degradation where traffic exceeds the PPS limit of the Exinda Appliance, it is recommended that the Clustering feature is disabled.
CauseExinda Appliances when configured in Cluster and high availability network architecture, every packet received on the active Exinda is sent to secondary Exinda as well for QOS and reporting purposes. The increased amount of Packets cause the performance degradation.
WorkaroundDisable clustering. Or Make changes to Cluster configuration, for this open case with Exinda support team.
ResolutionEngineering team will address this issue in a codified fix. ETA is not available at the moment.
For Exinda TAC Engineer, if the customer still wants to use clustering, Recommend below changes to cluster appliances.
Step1) Disable clustering on both the appliances master and slave Exinda. Install the shell license
Step 2) Type the following in configuration mode for both Exindas.
no cluster sync optimizer
Verify that the setting has changed from enabled to disabled for optimizer using:
show cluster sync
Step 3) After implementing the above commands on both the clustered Exindas. Enable clustering on the Exindas.
Impact of this command:
This Disables sync. of packets for the optimizer. i.e. copying meta data for every packet so they can be replayed on the other Exinda so that policy/VC/circuit bandwidth is globally enforced. This means reports will not synchronize on the clustered Exindas but configuration will synchronize.