Skip to content

Evolutionary infrastructure backup

Evolutionary infrastructure backup is a strategy used in cloud infrastructure management, specifically for Kubernetes clusters, that prioritizes the preservation and transition of existing environments rather than in-place replacement or deletion.^[400-devops-05-cloud-provider-gcp.md]

Core Mechanism

Instead of upgrading a live node pool directly—which risks downtime or data loss—this method involves creating a secondary, temporary environment to absorb the workload.^[400-devops-05-cloud-provider-gcp.md] This allows the original infrastructure to be upgraded safely while services remain operational in the temporary space.^[400-devops-05-cloud-provider-gcp.md]

Operational Workflow

The process typically follows a sequence of creating a new environment, shifting traffic, upgrading the old environment, and then shifting traffic back:

  1. Preparation: Infrastructure tools are initialized and configured to ensure the local state matches the cloud provider's current state^[400-devops-05-cloud-provider-gcp.md].
  2. Temporary Node Creation: A new node pool (often named temp or similar) is provisioned using infrastructure-as-code (e.g., Terraform) to serve as a backup target^[400-devops-05-cloud-provider-gcp.md].
  3. Workload Migration: Application configurations are updated to direct deployment away from the original nodes and towards the temporary nodes^[400-devops-05-cloud-provider-gcp.md]. This may involve a rolling update strategy (e.g., rotating Kafka nodes at intervals) to minimize disruption^[400-devops-05-cloud-provider-gcp.md].
  4. Primary Upgrade: With the workload safely shifted to the temporary pool, the original node pool is upgraded to the new version^[400-devops-05-cloud-provider-gcp.md].
  5. Restoration: Workloads are migrated back to the now-upgraded original node pool^[400-devops-05-cloud-provider-gcp.md], and the temporary resources are decommissioned or preserved for future cycles.

Sources

  • 400-devops-05-cloud-provider-gcp.md