Can't build notebook images locally

[kimwnasptd]

/kind bug

What steps did you take and what happened:
After trying to debug why the image builds failed in https://github.com/kubeflow/kubeflow/actions/runs/9216437170/job/25407648295 I tried to build the images locally to try and reproduce the error.

But I can't build the images locally and get the following error:

$ make docker-build-multi-arch--jupyter-tensorflow
make docker-build-multi-arch -C jupyter-tensorflow
make[1]: Entering directory '/home/ubuntu/Code/git/kubeflow/components/example-notebook-servers/jupyter-tensorflow'



------------------------------------------------------------------------------
Building 'jupyter-tensorflow' image for 'linux/amd64'...
------------------------------------------------------------------------------

docker buildx build --load --platform linux/amd64 --build-arg BASE_IMG=docker.io/kubeflownotebookswg/jupyter:sha-8869f221655b1de52b8298d29170ee0d06a46500 --progress=plain --cache-from type=registry,ref=ghcr.io/kubeflow/kubeflow/notebook-servers/build-cache:jupyter-tensorflow --tag "docker.io/kubeflownotebookswg/jupyter-tensorflow:sha-8869f221655b1de52b8298d29170ee0d06a46500" -f Dockerfile .
#0 building with "silly_ardinghelli" instance using docker-container driver

#1 [internal] load build definition from Dockerfile
#1 transferring dockerfile: 501B done
#1 DONE 0.0s

#2 [internal] load metadata for docker.io/kubeflownotebookswg/jupyter:sha-8869f221655b1de52b8298d29170ee0d06a46500
#2 ERROR: docker.io/kubeflownotebookswg/jupyter:sha-8869f221655b1de52b8298d29170ee0d06a46500: not found
------
 > [internal] load metadata for docker.io/kubeflownotebookswg/jupyter:sha-8869f221655b1de52b8298d29170ee0d06a46500:
------
Dockerfile:6
--------------------
   4 |
   5 |     ARG BASE_IMG=<jupyter>
   6 | >>> FROM $BASE_IMG
   7 |
   8 |     # args - software versions
--------------------
ERROR: failed to solve: docker.io/kubeflownotebookswg/jupyter:sha-8869f221655b1de52b8298d29170ee0d06a46500: failed to resolve source metadata for docker.io/kubeflownotebookswg/jupyter:sha-8869f221655b1de52b8298d29170ee0d06a46500: docker.io/kubeflownotebookswg/jupyter:sha-8869f221655b1de52b8298d29170ee0d06a46500: not found
make[1]: *** [../common.mk:88: docker-build-multi-arch] Error 1
make[1]: Leaving directory '/home/ubuntu/Code/git/kubeflow/components/example-notebook-servers/jupyter-tensorflow'
make: *** [Makefile:41: docker-build-multi-arch--jupyter-tensorflow] Error 2

I then tried to build all the images, in case I'm missing something with the dependencies, but got a similar error for the codeserver image

$ make docker-build-multi-arch
make docker-build-multi-arch -C base
make[1]: Entering directory '/home/ubuntu/Code/git/kubeflow/components/example-notebook-servers/base'



------------------------------------------------------------------------------
Building 'base' image for 'linux/amd64'...
------------------------------------------------------------------------------

...
...

------------------------------------------------------------------------------
Building 'codeserver' image for 'linux/amd64'...
------------------------------------------------------------------------------

docker buildx build --load --platform linux/amd64 --build-arg BASE_IMG=docker.io/kubeflownotebookswg/base:sha-8869f221655b1de52b8298d29170ee0d06a46500 --progress=plain --cache-from type=registry,ref=ghcr.io/kubeflow/kubeflow/notebook-servers/build-cache:codeserver --tag "docker.io/kubeflownotebookswg/codeserver:sha-8869f221655b1de52b8298d29170ee0d06a46500" -f Dockerfile .
#0 building with "silly_ardinghelli" instance using docker-container driver

#1 [internal] load build definition from Dockerfile
#1 transferring dockerfile: 706B done
#1 DONE 0.0s

#2 [internal] load metadata for docker.io/kubeflownotebookswg/base:sha-8869f221655b1de52b8298d29170ee0d06a46500
#2 ERROR: docker.io/kubeflownotebookswg/base:sha-8869f221655b1de52b8298d29170ee0d06a46500: not found
------
 > [internal] load metadata for docker.io/kubeflownotebookswg/base:sha-8869f221655b1de52b8298d29170ee0d06a46500:
------
Dockerfile:6
--------------------
   4 |
   5 |     ARG BASE_IMG=<base>
   6 | >>> FROM $BASE_IMG
   7 |
   8 |     ARG TARGETARCH
--------------------
ERROR: failed to solve: docker.io/kubeflownotebookswg/base:sha-8869f221655b1de52b8298d29170ee0d06a46500: failed to resolve source metadata for docker.io/kubeflownotebookswg/base:sha-8869f221655b1de52b8298d29170ee0d06a46500: docker.io/kubeflownotebookswg/base:sha-8869f221655b1de52b8298d29170ee0d06a46500: not found
make[1]: *** [../common.mk:88: docker-build-multi-arch] Error 1
make[1]: Leaving directory '/home/ubuntu/Code/git/kubeflow/components/example-notebook-servers/codeserver'
make: *** [Makefile:41: docker-build-multi-arch--codeserver] Error 2

What did you expect to happen:
To be able to build the images locally

Environment:

• Kubeflow version: master (pre 1.9 rc0)

cc @thesuperzapper

[thesuperzapper]

@kimwnasptd you are running the wrong make command.

The correct command to build (with dependencies) is to cd into the folder of the image you want to build and run one of the following:

• Local Arch: make docker-build-dep
• Multi Arch: make docker-build-multi-arch-dep

You can also build ALL images by running the following commands from the ./components/example-notebook-servers/ folder:

• Local Arch: make docker-build
• Multi-Arch: make docker-build-multi-arch

---

Also, if you want to push to a test registry, you can run export REGISTRY="docker.io/MY_DOCKER_ACCOUNT" (or any other registry) before running the "build" commands, then you can run the corresponding "push" command:

• Single Image (+ Deps)
  • Local Arch: make docker-push-dep
  • Multi-Arch: make docker-build-push-multi-arch-dep (note, this also builds)
• All Images:
  • Local Arch: make docker-push
  • Multi-Arch: make docker-build-multi-arch (note, this also builds)

[thesuperzapper]

Closing, as #7590 added a local build guide.

See https://github.com/kubeflow/kubeflow/tree/master/components/example-notebook-servers#build-images