Fix kubeflowkatib/mxnet-mnist image

[tenzen-y]

What this PR does / why we need it:

It seems that mxnet/python:latest_cpu_native_py3 image has been deleted from upstream dockerhub.
As a result, our suite test can no longer be run as shown below, so I changed the base image from mxnet/python:latest_cpu_native_py3 to python:3.9.

error building image: GET https://index.docker.io/v2/mxnet/python/manifests/latest_cpu_native_py3: MANIFEST_UNKNOWN: manifest unknown; unknown tag=latest_cpu_native_py3

Argo Workflow Logs

Also, I have used ytenzen/katib-mxnet-mnist:fix-mxnet-image to check the operation of the new image.

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
Fixes #

Checklist:

• Docs included if any changes are user facing

[tenzen-y]

When I use mxnet/python:2.0.0beta1_cpu_py3 as the base image, I have faced on below errors.
So, it might be better to modify examples for mxnet in the future.

Downloading data/train-labels-idx1-ubyte.gz from https://repo.mxnet.io/gluon/dataset/mnist/train-labels-idx1-ubyte.gz...
Downloading data/train-images-idx3-ubyte.gz from https://repo.mxnet.io/gluon/dataset/mnist/train-images-idx3-ubyte.gz...
Downloading data/t10k-labels-idx1-ubyte.gz from https://repo.mxnet.io/gluon/dataset/mnist/t10k-labels-idx1-ubyte.gz...
Downloading data/t10k-images-idx3-ubyte.gz from https://repo.mxnet.io/gluon/dataset/mnist/t10k-images-idx3-ubyte.gz...
[17:14:13] ../src/storage/storage.cc:202: Using Pooled (Naive) StorageManager for CPU
Traceback (most recent call last):
  File "/opt/mxnet-mnist/mnist.py", line 86, in <module>
    fit.fit(args, sym, get_mnist_iter)
  File "/opt/mxnet-mnist/common/fit.py", line 228, in fit
    model = mx.mod.Module(
AttributeError: module 'mxnet' has no attribute 'mod'

[coveralls]

Coverage increased (+0.2%) to 73.978% when pulling 0f281d1 on tenzen-y:fix-mxnet-image into ea23e71 on kubeflow:master.

[tenzen-y]

Also, this PR blocks #1865 and #1833.

[tenzen-y]

/cc @andreyvelich @johnugeorge @gaocegege

[tenzen-y]

It seems to fail to deploy katib in Charmed Katib/ Test.

Traceback (most recent call last):
File "/usr/local/lib/python3.8/dist-packages/charmcraft/charm_builder.py", line 361, in
main()
File "/usr/local/lib/python3.8/dist-packages/charmcraft/charm_builder.py", line 357, in main
builder.build_charm()
File "/usr/local/lib/python3.8/dist-packages/charmcraft/charm_builder.py", line 93, in build_charm
self.handle_dependencies()
File "/usr/local/lib/python3.8/dist-packages/charmcraft/charm_builder.py", line 248, in handle_dependencies
_process_run(cmd)
File "/usr/local/lib/python3.8/dist-packages/charmcraft/charm_builder.py", line 299, in _process_run
raise CommandError(f"Subprocess command {cmd} execution failed with retcode {retcode}")
charmcraft.cmdbase.CommandError: Subprocess command ['/home/runner/work/katib/katib/operators/katib-ui/build/parts/charm/build/staging-venv/bin/pip3', 'install', '--upgrade', '--no-binary', ':all:', '--requirement=requirements.txt'] execution failed with retcode 1
Parts processing error: Failed to run the build script for part 'charm'. (full execution logs in '/tmp/charmcraft-log-xp6383qh')
Error: Error running subcommand charmcraft pack -p ./katib-ui --destructive-mode: exit status: 1
Error: Process completed with exit code 1.

https://github.com/kubeflow/katib/runs/6475581693?check_suite_focus=true#step:6:372

Could you help to resolve this problem? @DomFleischmann @knkski @ca-scribner

[gaocegege]

/lgtm

[ca-scribner]

I don't recognise this error, but I'll get someone to take a look. Should be a quick fix and likely is related to some tooling and not actually broken by this PR.

[ca-scribner]

Waiting on CI to confirm but this should be fixed by #1867.

[tenzen-y]

Waiting on CI to confirm but this should be fixed by #1867.

Thank you for letting me know @ca-scribner.

[johnugeorge]

Please rebase !

[tenzen-y]

Please rebase !

@johnugeorge I have rebased this branch.

[johnugeorge]

Thanks @tenzen-y

/lgtm
/approve

[google-oss-prow]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: johnugeorge, tenzen-y

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [johnugeorge]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment