Omniverse Content Cache (OVCC)

OVCC (formerly USD Content Cache, UCC) provides a reverse proxy cache between object storage and Cloud Function clients. It caches USD (Universal Scene Description) content to accelerate scene loading and reduce bandwidth costs.

The Helm chart is published as ovcontentcache on NGC; OVCC is the short product name used in prose.

Overview

OVCC intercepts requests for USD content from object storage sources (S3, Azure Blob Storage, NVIDIA Omniverse Nucleus) and serves cached content when available. On cache hits, content is served directly from persistent storage. On cache misses, OVCC fetches content from the upstream source, caches it locally, and serves it to the client. This reduces egress costs from cloud storage providers and improves scene load times.

When to Use OVCC

• Working with USD scenes stored in object storage

• Want to reduce bandwidth usage and egress costs

• Need faster scene loading times

Documentation

• Configuration Guide - Helm values, scaling, storage, and cache settings

• Deployment Guide - Installing OVCC using Helm

• TLS Configuration - Enable encrypted communication

• Runbooks - Troubleshooting common operational issues

• OVCC: Configure
  • Base Configuration
  • Release and Service Names
  • Complete Configuration Reference
  • 1. Provision and Scale
  • 2. Storage Configuration
  • Example Kubernetes StorageClass for AWS
  • 3. Cache Configuration
  • 4. Telemetry
  • Summary
• OVCC: Deployment
  • Configuration
  • Deployment Steps
  • 1. Configure Helm
  • 2. Prepare Namespace
  • 3. Create Required Secrets
  • 4. Install Helm Chart
  • 5. Verify Deployment
  • Upgrading from USD Content Cache (UCC)
  • Uninstall Instructions
  • Summary
• OVCC: TLS Configuration
  • 1. Create the TLS Secret
  • Self-Signed Certificate (Reference Only)
  • 2. Enable TLS in values.yaml
  • 3. Complete Example Configuration
  • 4. Apply Configuration Changes
  • 5. Verify TLS Configuration
  • Summary

Runbooks

• OVCC Runbooks
  • Available Runbooks
• OVCC: Connection Saturation and High Response Times
  • Overview
  • Symptoms and Detection Signals
    • Visible Symptoms
    • Log Messages
      • Connection Refused
      • Timeout Errors
    • Metric Signals
  • Root Cause Analysis
    • Known Causes
      • Undersized Worker Connection Limits
      • Insufficient CPU Cores
      • Too Few Replicas
    • Other Possible Causes
  • Troubleshooting Steps
    • Diagnostic Steps for Known Root Causes
      • 1. Monitor Connection Metrics and Identify Saturation
      • 2. Increase Worker Connection Limits
      • 3. Scale OVCC Replicas to Distribute Load
      • 4. Increase CPU Allocation for More Worker Processes
    • Other Diagnostic Actions
  • Prevention
    • Proactive Monitoring
    • Configuration Best Practices
    • Capacity Planning
• OVCC: Data Disk Bandwidth Bottlenecks
  • Overview
  • Symptoms and Detection Signals
    • Visible Symptoms
    • Metric Signals
  • Root Cause Analysis
    • Known Causes
      • Insufficient Storage Class Tier
      • PVC Size Too Small for Bandwidth Scaling
    • Other Possible Causes
  • Troubleshooting Steps
    • Diagnostic Steps for Known Root Causes
      • 1. Monitor Disk Bandwidth Usage and Identify Saturation
      • 2. Upgrade to Burst-Capable Storage Class
      • 3. Increase PVC Size to Scale Bandwidth
  • Prevention
    • Proactive Monitoring
    • Configuration Best Practices
• OVCC: Metadata Cache Undersizing
  • Overview
  • Symptoms and Detection Signals
    • Visible Symptoms
    • Cache Miss Patterns
    • Metric Signals
  • Root Cause Analysis
    • Known Causes
      • Default Metadata Cache Size Insufficient
      • Configuration Applied to Wrong Helm Values Section
    • Other Possible Causes
  • Troubleshooting Steps
    • Diagnostic Steps
      • 1. Verify Configuration Was Applied Correctly
      • 2. Monitor Cache Miss Rates and Decide If Resize Is Needed
      • 3. Increase Metadata Cache Size
  • Prevention
    • Proactive Monitoring
    • Configuration Best Practices
    • Capacity Planning
• OVCC: Network Bandwidth Saturation
  • Overview
  • Symptoms and Detection Signals
    • Visible Symptoms
    • Metric Signals
  • Root Cause Analysis
    • Known Causes
      • Too Few OVCC Replicas
      • Uneven Load Distribution
    • Other Possible Causes
  • Troubleshooting Steps
    • Diagnostic Steps
      • 1. Confirm Network is the Bottleneck (Not Disk or Cache)
      • 2. Scale OVCC Replicas to Distribute Network Load
      • 3. Verify Load Distribution Across Replicas
      • 4. Check Pod Placement in Same Availability Zone
  • Prevention
    • Proactive Monitoring
    • Configuration Best Practices
• OVCC: Upstream S3 Connection Spikes and High Connect Time
  • Overview
  • Symptoms and Detection Signals
    • Visible Symptoms
    • Log Messages
      • S3 503 Service Unavailable
    • Metric Signals
  • Root Cause Analysis
    • Known Causes
      • No Connection Reuse to S3
      • High Cache MISS Rate Driving S3 Requests
    • Other Possible Causes
  • Troubleshooting Steps
    • Diagnostic Steps
      • 1. Monitor Upstream Connection Count and Connect Time
      • 2. Reduce Cache MISS Rate to Lower S3 Request Volume
      • 3. Monitor Cloud Provider SNAT Allocation
  • Prevention
    • Proactive Monitoring
    • Mitigation Strategies
    • Future Improvements