Interpreting Vertica Diagnostic Results (vcpuperf and vioperf) for DX NetOps Performance Management Readiness

Products

• Network Observability
• CA Performance Management

Issue/Introduction

This article provides guidance on interpreting the output of Vertica diagnostic scripts (vcpuperf and vioperf). Understanding these results is critical for ensuring your server environment meets the performance requirements for a successful DX NetOps Performance Management Data Repository installation.

Environment

*   DX NetOps Performance Management (All Supported Versions)

*   Vertica Analytic Database (All Supported Versions)

Resolution

Diagnostic Procedures:

1. Reviewing vcpuperf output: This test measures CPU latency. Ensure the results consistently show low latency benchmarks, as high spikes indicate potential performance bottlenecks during Vertica query processing.
2. Reviewing vioperf output: This test measures Disk I/O throughput. Compare the results against the minimum required throughput specified in the DX NetOps Performance Management documentation for your cluster size.
3. Identifying Anomalies: If vioperf results fall below the required thresholds, verify disk controller settings and storage subsystem alignment before proceeding with installation.

Best Practices:

• Always run diagnostics on all nodes in the cluster before attempting installation.
• If you detect inconsistencies between nodes, investigate hardware or hypervisor resource contention.
• If installation is required despite sub-optimal results, document the findings in your implementation plan and consult with Broadcom Support for risk assessment.

Additional Information

Run Data Repository Diagnostic Utilities