Skip to Main Content
Reducing Container Pull and Startup Time on OCI VMs with FSS/Lustre Caching

About This Workshop

Youtube Video

About This Workshop
This workshop provides a step-by-step guide to reducing container image pull and startup time on OCI VMs by introducing a shared, network-attached cache for container layers using OCI File Storage Service (FSS) and Lustre. You will provision the required OCI infrastructure with Terraform (VCN, compute, and shared storage), then use a cloud-init–driven VM bootstrap to install and configure Podman with additionalimagestores pointing to the shared cache paths.

You will run a repeatable benchmark workflow to compare cold vs warm container pulls and container starts, validate that cached layers are being reused across runs, and capture concrete timing results that demonstrate the impact of shared caching on startup latency. The lab progresses incrementally—starting with FSS for simplicity, then extending the same approach to Lustre for higher-performance scenarios—so participants leave with a practical reference implementation and a clear decision framework for when to use FSS vs Lustre for large-scale deployments.

Workshop Info

2 hours
  • Lab 1: Provision of the necessary resources
  • Lab 2: Explore FSS and Lustre benchmarking scenarios
  • Lab 3: Cleanup
  • An Oracle CLoud account
  • Sufficient privileges to create/manage OCI resources used in the lab (VCN, Compute, FSS, Lustre, IAM policies related)
  • Basic understanding of containers and VM-based runtimes (images, layers, pull vs run/start) and basic Linux CLI skills.
  • Access to a command-line environment such as OCI Cloud Shell or a local seup with:
    • terraform
    • oci CLI configured for your tenancy
    • ssh (and an SSH keypair)
  • Ability to connect to the provisioned VM (SSH allowed from your client IP/Cloud Shell) 

Other Workshops you might like