| commit | 0e3262c1220f8fecf91a5142b18070f6c3fe49e6 | [log] [tgz] |
|---|---|---|
| author | Jakob Buchgraber <buchgr@google.com> | Fri Feb 16 21:41:03 2018 +0100 |
| committer | Jakob Buchgraber <buchgr@google.com> | Fri Feb 16 21:41:03 2018 +0100 |
| tree | c289009a7d66d14947f58510171b070e8a7c25a4 | |
| parent | 3d52d606f4625c47bb4cf7a20bc9463536ae6336 [diff] |
add tensorflow configuration
This workspace contains the setup for the continuous integration system of Bazel. This setup is based on docker images built by bazel.
If you are a user of the CI system, you might be interested in the following document:
Make sure you have a Bazel installed with a recent enough version of it. Also make sure gcloud and docker are correctly configured on your machine. Only docker version 1.10 or later is supported.
Finally, our docker rules needs authentication which is configured with the credential helper:
gcloud components install docker-credential-gcr export DOCKER_CONFIG="$(docker-credential-gcr configure-docker | sed -E 's|/config.json .*$||')"
You might want to permanently set your DOCKER_CONFIG environment, e.g. in your ~/.bash_profile:
echo "export DOCKER_CONFIG='${DOCKER_CONFIG}'" >> ~/.bash_profile
More documentation:
init.sh: initializes the whole CI platform. This may delete VMs and do other irreversible changes, so handle with care.vm.sh: lets you control the machines (e.g. start/stop them, create/delete/reimage them), including the Jenkins controller and the executor nodes.