tensorflow: Failed to get device properties, error code: 30

Describe the current behavior I commented under issue #26255 but the original poster closed the issue as his problem was solved by updating to tensorflow 2.

I am opening a new issue because updating to the pre-release is not an option and I have no way to even trap this error to try to handle it, plus it is an unknown error code so no hint as to how to proceed.

Unknown error and failure to initialize GPU.

2019-04-19 13:16:22.838705: E tensorflow/core/grappler/clusters/utils.cc:83] Failed to get device properties, error code: 30
Failed to initialize GPU device #0: unknown error

My configuration: Windows 10 Home Tensorflow 1.13.1 Python 3.5 GTX 1060 Mobile Max-Q

It doesn’t happen every time I run my program. I have localized it to running load_model from keras, before reaching that point I have imported tensorflow and verified that gpu is available.

  if tf.test.is_gpu_available():
        logger.debug("GPU is available")

Is there a way to catch this error or check for it and recover/attempt to reinitialize?

Thanks Describe the expected behavior Report the actual error, provide mechanism to catch the exception and handle it.

Code to reproduce the issue The failure is intermittent

Other info / logs Include any logs or source code that would be helpful to diagnose the problem. If including tracebacks, please include the full traceback. Large logs and files should be attached.