Serving: error loading package '': Extension file 'third_party/clang_toolchain/cc_configure_clang.bzl' has errors.

[zhixiaochuan12]

I am running a tensorflow serving container referring to this , all the previous steps are good, but in the last block I met some problems:

git clone --recurse-submodules https://github.com/tensorflow/serving
cd serving/tensorflow
./configure
cd ..
bazel test tensorflow_serving/...

I did not do the ./configure line because that there is not a serving/tensorflow dir and I saw someone in another issue said there is no need to do that any more. So I directly does to the last linebazel test tensorflow_serving/..., but the error occurs:

root@15bb1c2766e3:/serving# bazel test tensorflow_serving/...
ERROR: /root/.cache/bazel/_bazel_root/f8d1071c69ea316497c31e40fe01608c/external/org_tensorflow/third_party/clang_toolchain/cc_configure_clang.bzl:3:1: file '@bazel_tools//tools/cpp:cc_configure.bzl' does not contain symbol 'cc_autoconf_impl'.
ERROR: error loading package '': Extension file 'third_party/clang_toolchain/cc_configure_clang.bzl' has errors.
ERROR: error loading package '': Extension file 'third_party/clang_toolchain/cc_configure_clang.bzl' has errors.
INFO: Elapsed time: 0.107s
ERROR: Couldn't start the build. Unable to run tests.

And in my container, the bazel version is 0.5.4:

root@15bb1c2766e3:/serving# bazel version
Build label: 0.5.4
Build target: bazel-out/local-fastbuild/bin/src/main/java/com/google/devtools/build/lib/bazel/BazelServer_deploy.jar
Build time: Fri Aug 25 10:00:00 2017 (1503655200)
Build timestamp: 1503655200
Build timestamp as int: 1503655200

Below is my bazel.rc:

root@15bb1c2766e3:/serving/tools# cat bazel.rc
build:cuda --crosstool_top=@org_tensorflow//third_party/gpus/crosstool
build:cuda --define=using_cuda=true --define=using_cuda_nvcc=true

build --force_python=py3
build --python3_path=/usr/bin/python

build --action_env PYTHON_BIN_PATH="/usr/bin/python"

build --define PYTHON_BIN_PATH=/usr/bin/python

build --spawn_strategy=standalone --genrule_strategy=standalone
test --spawn_strategy=standalone --genrule_strategy=standalone
run --spawn_strategy=standalone --genrule_strategy=standalone

build --define=grpc_no_ares=true

# TODO(b/69809703): Remove once no longer required for TensorFlow to build.
build --copt=-DGEMMLOWP_ALLOW_SLOW_SCALAR_FALLBACK

I guess there may be something dismatch, but I can not find much information about the third_party/clang_toolchain.

[ghost]

I'm facing the same problem.

[guillaumegenthial]

possible merge (and fix) with #854.
TL;DR: change Dockerfile.devel BAZEL_VERSION to 0.12.0

[RodMayen]

So do I, same problem, and already tried suggestion by guillaumegenthial without succesfully results. Any other idea guys?

[AswathKiruba]

We could also update bazel directly in the docker container by

1. Download bazel--installer-linux-x86_64.sh from https://github.com/bazelbuild/bazel/releases location
2. chmod +x ./bazel--installer-linux-x86_64.sh
3. ./bazel--installer-linux-x86_64.sh

@RodMayen Are using the tensorflow serving master (1.6) branch? Also please do a bazel clean the before building.

[RodMayen]

@AswathKiruba thanks a lot for your answer, actually I was wrong, the suggestion by @guillaumegenthial was the way to go for me! thanks a lot guys

[zhixiaochuan12]

@guillaumegenthial Thanks, It is the issue.