Cut beta release of tf-operator for 1.4 release

[Jeffwan]

We notice a few issues recently

1. Gang scheduling issue with Volcano #1382
2. installation issue due to kustomize version or kubectl #1381
3. Testing on existing SDK
4. E2E testing for rest of the frameworks besides TF.

I will spend some time this weekend to address some of above issue before we cut beta release.

[Jeffwan]

SDK installation issue

1. Failed to create the job. but our e2e integration pass. it's probably my env's problem

tfjob_client = TFJobClient()
tfjob_client.create(tfjob, namespace=namespace)


    236             # model definition for request.
    237             obj_dict = {obj.attribute_map[attr]: getattr(obj, attr)
--> 238                         for attr, _ in six.iteritems(obj.openapi_types)
    239                         if getattr(obj, attr) is not None}
    240 

AttributeError: 'V1TFJob' object has no attribute 'openapi_types'

2. Installation issue

➜  python git:(sdk) ✗ python3 setup.py install --user
running install
error: can't combine user with prefix, exec_prefix/home, or install_(plat)base

works after adding `--prefix=`

➜  python git:(sdk) ✗ python3 setup.py install --user --prefix=
running install
running bdist_egg
.....
Using /usr/local/lib/python3.9/site-packages
Finished processing dependencies for kubeflow-tfjob==0.1.4

[Jeffwan]

Manually test all 4 frameworks, all work well

apiVersion: "kubeflow.org/v1"
kind: TFJob
metadata:
  name: test
  namespace: kubeflow
spec:
   tfReplicaSpecs:
     Worker:
       replicas: 2
       restartPolicy: OnFailure
       template:
         spec:
           containers:
             - name: tensorflow
               image: gcr.io/kubeflow-ci/tf-mnist-with-summaries:1.0
               command:
                 - "python"
                 - "/var/tf_mnist/mnist_with_summaries.py"
~

apiVersion: "kubeflow.org/v1"
kind: PyTorchJob
metadata:
  name: pytorch
  namespace: kubeflow
spec:
  pytorchReplicaSpecs:
    Master:
      replicas: 1
      restartPolicy: OnFailure
      template:
        spec:
          containers:
            - name: pytorch
              image: docker.io/kubeflowkatib/pytorch-mnist:v1beta1-45c5727
              imagePullPolicy: Always
              command:
                - "python3"
                - "/opt/pytorch-mnist/mnist.py"
                - "--epochs=1"
    Worker:
      replicas: 1
      restartPolicy: OnFailure
      template:
        spec:
          containers:
            - name: pytorch
              image: docker.io/kubeflowkatib/pytorch-mnist:v1beta1-45c5727
              imagePullPolicy: Always
              command:
                - "python3"
                - "/opt/pytorch-mnist/mnist.py"
                - "--epochs=1"
~

https://github.com/kubeflow/tf-operator/blob/master/examples/mxnet/mxjob_dist_v1.yaml
https://github.com/kubeflow/tf-operator/blob/master/examples/xgboost/xgboostjob.yaml

[Jeffwan]

I cut https://github.com/kubeflow/tf-operator/releases/tag/v1.3.0-rc.0 release. Official release would be discussed in the WG-training meeting.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.