We successfully migrated our reverse logistics customer from on-premises Rancher Kubernetes servers to Google Kubernetes Engine on GCP.
In this case, SCV successfully migrated from on-premises servers with rancher kubernetes to GCP servers on Google Kubernetes Engine (GKE) cloud for our reverse logistic customer.
From the custom on-premises servers at a datacenter in Washington, DC, the running operations of the client were migrated to the servers in Google Cloud. The servers originally ran on Linux, had Kubernetes installed on them, and the rancher interface managed the nodes for adding/modifying/deleting. In the final state, the servers were created on Google Cloud Platform and used the GKE interface for creating and adding nodes.
Many months of work led to a number of individual steps. As each team began to own their own service, a slow migration process started. First, the infrastructure team created the cluster and enabled a “bridge” on the network level between both clusters (we used Kubernetes inter-cluster configuration).
Second, each team modified ArgoCD parameters so that new clusters created a replica of their configuration on GKE. For each service, when the data was synchronized, the external and internal DNS was remapped to the new GCP replica also via ArgoCD parameters.
Lastly, for migrating the kafka bus-service, SCV switched on a specific deploy that started on a Sunday (the usual timing for the company). To accomplish this, the complete system was taken down while we changed the DNS resolution entries from the old to the new cluster (the Kubernetes Services changed internal IPs to real pods on same cluster), and the system was restarted.
SCV’s client could finally lower their base-cost as they went away from buying servers on an ON-PREM datacenter based in Virginia (paying base costs). They could also avoid intermittent network issues on the VLAN that added a 5% package loss on the custom-made Kubernetes Engine (used a Canal VLAN configuration), and switched to the same configuration on GKE, but fine tuned for a specific Hosted Linux Kernel. All this was successfully accomplished due to the careful planning and orchestration of the team members without data loss.
We built a platform connecting indigenous communities with nature stewardship projects, creating quantifiable biodiversity credits for ESG.
We'd love to hear about your project and ideas.
