Cloud Code support for ARM and mixed-architecture GKE clusters

This page outlines Cloud Code's support for deploying to ARM and mixed-architecture Google Kubernetes Engine clusters, as well as minikube clusters.

To learn more about ARM workloads, see Managing ARM workloads.

Warning: The information on this page applies to Skaffold v2. If you use managed dependencies, you'll already have v2. If you're still using Skaffold v1 which is no longer actively maintained, the support mentioned on this page may not work as described.

Note: minikube does not emulate a platform that isn't your client's platform. For example, you can only build an ARM image for minikube and run it if you are on an ARM client.

ARM and mixed-architecture support tables

Refer to the following table for support based on the workstation you're developing on:

x86 workstations

This table refers to developers using x86 workstations which are deploying to either ARM or mixed-architecture GKE clusters. All of the actions listed under Actions are supported for deployments to both ARM and mixed-architecture GKE clusters, except for locally building an ARM image using Buildpacks:

Actions	ARM GKE cluster	Mixed-architecture GKE cluster
Use Cloud Code to locally build an ARM image using Jib
Use Cloud Code to locally build an ARM image using Dockerfiles
Use Cloud Code to locally build an ARM image using Buildpacks
Deploy your ARM image without manual intervention and run or debug your application

Note: Docker Engine on Linux does not come with QEMU emulators installed by default, which is required by Skaffold's use of Docker BuildX for cross-platform builds. To setup QEMU for Docker builds, see the instructions in Docker BuildX.

ARM workstations

This table refers to developers using ARM workstations which are deploying to either ARM or mixed-architecture GKE clusters. All of the actions listed under Actions are supported for deployments to both ARM and mixed-architecture GKE clusters, as well as minikube clusters, except for locally building an ARM image using Buildpacks:

Actions	ARM GKE cluster	Mixed-architecture GKE cluster	minikube cluster
Use Cloud Code to locally build an ARM image using Jib
Use Cloud Code to locally build an ARM image using Dockerfiles
Use Cloud Code to locally build an ARM image using Buildpacks
Deploy your ARM image without manual intervention and run or debug your application

To learn more about how you can specify the platform you want to build your images for and override Skaffold's automatic image building logic, see Deploy your applications to ARM or mixed-architecture GKE clusters

Get Support

To send feedback, report issues on GitHub, or ask a question on Stack Overflow.