notebook_setup() error when running mnist_gcp.ipynb

[Linchin]

I'm trying to follow the mnist_gcp.ipynb notebook to make sure my newly installed kubeflow is working properly. When I executed the second cell, an error popped up. Here is the cell:

import notebook_setup
reload(notebook_setup)
notebook_setup.notebook_setup()

The error message looks like this:

pip installing requirements.txt
Checkout kubeflow/tf-operator @9238906
Configure docker credentials
---------------------------------------------------------------------------
FileNotFoundError                         Traceback (most recent call last)
<ipython-input-6-b1e5ce107a08> in <module>
      1 import notebook_setup
      2 reload(notebook_setup)
----> 3 notebook_setup.notebook_setup()

~/git_kubeflow-examples/mnist/notebook_setup.py in notebook_setup(platform)
     46 
     47   if platform == 'gcp':
---> 48     setup_gcp()
     49 
     50   # Force a reload of kubeflow; since kubeflow is a multi namespace module

~/git_kubeflow-examples/mnist/notebook_setup.py in setup_gcp()
     55 def setup_gcp():
     56   logging.info("Configure docker credentials")
---> 57   subprocess.check_call(["gcloud", "auth", "configure-docker", "--quiet"])
     58   if os.getenv("GOOGLE_APPLICATION_CREDENTIALS"):
     59     logging.info("Activating service account")

/opt/conda/lib/python3.8/subprocess.py in check_call(*popenargs, **kwargs)
    357     check_call(["ls", "-l"])
    358     """
--> 359     retcode = call(*popenargs, **kwargs)
    360     if retcode:
    361         cmd = kwargs.get("args")

/opt/conda/lib/python3.8/subprocess.py in call(timeout, *popenargs, **kwargs)
    338     retcode = call(["ls", "-l"])
    339     """
--> 340     with Popen(*popenargs, **kwargs) as p:
    341         try:
    342             return p.wait(timeout=timeout)

/opt/conda/lib/python3.8/subprocess.py in __init__(self, args, bufsize, executable, stdin, stdout, stderr, preexec_fn, close_fds, shell, cwd, env, universal_newlines, startupinfo, creationflags, restore_signals, start_new_session, pass_fds, encoding, errors, text)
    856                             encoding=encoding, errors=errors)
    857 
--> 858             self._execute_child(args, executable, preexec_fn, close_fds,
    859                                 pass_fds, cwd, env,
    860                                 startupinfo, creationflags, shell,

/opt/conda/lib/python3.8/subprocess.py in _execute_child(self, args, executable, preexec_fn, close_fds, pass_fds, cwd, env, startupinfo, creationflags, shell, p2cread, p2cwrite, c2pread, c2pwrite, errread, errwrite, restore_signals, start_new_session)
   1702                     if errno_num != 0:
   1703                         err_msg = os.strerror(errno_num)
-> 1704                     raise child_exception_type(errno_num, err_msg, err_filename)
   1705                 raise child_exception_type(err_msg)
   1706 

FileNotFoundError: [Errno 2] No such file or directory: 'gcloud'

Any idea why this is happening? Am I missing something?

[js-ts]

Hi @Linchin , which version of kubeflow have you installed , if you're using kubeflow <v1.4 you should use this notebook https://github.com/kubeflow/pipelines/blob/master/samples/contrib/kubeflow-e2e-mnist/kubeflow-e2e-mnist.ipynb
this notebook is used by users, distrubutions to see if their kubeflow cluster is working if you're trying out the v1.5 version with kserve v0.7
you should use https://github.com/js-ts/pipelines/blob/master/samples/contrib/kubeflow-e2e-mnist/kubeflow-e2e-mnist.ipynb as it uses kserve component instead of the old kfserving <v0.6.1

[gdupret]

notebook_setup is enable to install the dependencies, probably because TF1.5 is not available anymore on GCS