Best Practices for Configuring NetMRI Virtual Appliances

This section outlines the best practices that have been found specific to NetMRI virtual appliance installation on a VMware server:

• AMD Processors

Issue: Hardware platforms using AMD processors may present performance bottlenecks to VMware running NetMRI, in certain cases.

Recommendation: Avoid AMD processor hardware if at all possible. If this is not feasible, pay attention to the results of the platform benchmark and resolve any issues before installing NetMRI.

• Disk Performance

Issue: In high-performance applications, such as high license count deployments, NetMRI can be I/O intensive. If multiple applications are configured to access the same physical disk, NetMRI can experience significant read/write delays that might degrade its

performance.

Recommendation: Provision NetMRI to use dedicated disks/spindles.

• Processors with Hardware Virtualization Support

Issue: NetMRI benefits greatly from the hardware virtualization support included in the latest generation processors. The performance of older generations of processors may not be suitable for high-performance applications, such as high license count deployments.

Recommendation: Run NetMRI on the latest generation of processors having hardware virtualization support. Intel i7 based servers have been shown to be particularly effective.

• Server Performance Monitoring

Issue: NetMRI is a high-performance network analysis system that can be different than the enterprise applications you regularly monitor. It may use CPUs intensely for extended periods of time and may exceed the CPU thresholds currently set on the server

performance monitoring applications.

Recommendation: If monitoring is set up for your VMware server, you may need to raise or eliminate the alert threshold values for NetMRI to eliminate unnecessary alarms.

• vCPU Allocation

Issue: Due to context-switching overhead, allocating more vCPUs to NetMRI may hurt performance.

Recommendation: Adhere to the recommended vCPU allocation parameters listed in Adding vCPUs. If the system performance is suffering, check the CPU Ready State times. If times are high, try reducing the number of vCPUs and see if that helps. If times

are low, try increasing vCPUs. For additional information relative to vCPU allocation, refer to the documentation on Performance Best Practices for VMware vSphere.

• Memory Allocation

Issue: Additional memory helps almost any application. If memory can be increased for the NetMRI virtual appliance without over-subscribing, that will help performance.

Recommendation: It is recommended you adhere to at least the recommended memory allocation parameters set forth in Adding Memory.