Ignite tests fail with python 3.4 and 3.6

[yongtang]

While running tests on Travis CI, it looks like ignite tests //tensorflow_io/ignite:ignite_py_test fail with python 3.4:
https://travis-ci.org/tensorflow/io/jobs/467306695

Python 3.6 tests also fail with //tensorflow_io/ignite:ignite_py_test
https://travis-ci.org/tensorflow/io/jobs/467306697

It may worth to take a look.

[yongtang]

/cc @dmitrievanthony The CI failure is flaky so I am wondering if the issue could be that //tensorflow_io/ignite:ignite_py_test and //tensorflow_io/ignite:igfs_py_test could not run concurrently?

[dmitrievanthony]

Thanks, @yongtang. These tests can run concurrently, I think the root cause is another. Actually, when I start Ignite in Docker container I can't easily understand outside if Ignite is ready or not and I just wait for 5 seconds. Looks like 5 seconds waiting is not the best choice.

Let me think a bit about it, I'll check if it's possible to ensure that Ignite is ready without waiting. Meanwhile, if this fails are frequent I can increase the waiting period.

[dmitrievanthony]

BTW, do you know if it's possible to download raw log from our Travis without special characters? Currently I see the raw log like this:

travis_fold:start:worker_info
�[0K�[33;1mWorker information�[0m

[yongtang]

Thanks @dmitrievanthony for looking into the issue. I created a PR #35 which enables ignite test, and also --test_output=errors in bazel command so that the errors could be shown in the Travis logs.

From Travis CI it looks like the following error(in python 2.7) will show up:

======================================================================
ERROR: test_ignite_dataset_with_plain_client (__main__.IgniteDatasetTest)
Test Ignite Dataset with plain client.
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/travis/.cache/bazel/_bazel_travis/50a9fc5d0e0cf7fe188c59f2e8fe814c/sandbox/linux-sandbox/854/execroot/__main__/bazel-out/k8-fastbuild/bin/tensorflow_io/ignite/ignite_py_test.runfiles/__main__/tensorflow_io/ignite/python/tests/ignite_dataset_test.py", line 47, in test_ignite_dataset_with_plain_client
    ds = IgniteDataset(cache_name="SQL_PUBLIC_TEST_CACHE", port=42300)
  File "/home/travis/.cache/bazel/_bazel_travis/50a9fc5d0e0cf7fe188c59f2e8fe814c/sandbox/linux-sandbox/854/execroot/__main__/bazel-out/k8-fastbuild/bin/tensorflow_io/ignite/ignite_py_test.runfiles/__main__/tensorflow_io/ignite/python/ops/ignite_dataset_ops.py", line 745, in __init__
    self.cache_type = client.get_cache_type(cache_name)
  File "/home/travis/.cache/bazel/_bazel_travis/50a9fc5d0e0cf7fe188c59f2e8fe814c/sandbox/linux-sandbox/854/execroot/__main__/bazel-out/k8-fastbuild/bin/tensorflow_io/ignite/ignite_py_test.runfiles/__main__/tensorflow_io/ignite/python/ops/ignite_dataset_ops.py", line 439, in get_cache_type
    "Scan Query Error [status=%s, message='%s']" % (status, err_msg))
RuntimeError: Scan Query Error [status=1000, message='Cache does not exist [cacheId= -919715654]']

----------------------------------------------------------------------
Ran 2 tests in 0.097s

FAILED (errors=1)
================================================================================

[yongtang]

With the PR #37 merged, this issue has been fix, so closing.