Titan tensorflow container fails to build about container-recipes

[AdamSimpson]

This appears related to bazelbuild/bazel#4483 . Due to the cuda driver restricting the version of tensorflow and the version of tensorflow restricting the version of bazel i'll see what I can do about rolling back the jdk version used.

[markcoletti]

While you explore rolling back, I'll explore rolling bazel forward to its more recent version, 0.11.1.

[markcoletti]

bazel now builds, but it does whine:

WARNING: /tmp/bazel_Sn65u98e/out/external/bazel_tools/WORKSPACE:1: Workspace name in /tmp/bazel_Sn65u98e/out/external/bazel_tools/WORKSPACE (@io_bazel) does not match the name given in the repository's definition (@bazel_tools); this will cause a build error in future versions

Build is proceeding. I'll close this out when and if it finishes.

[markcoletti]

And, this is the latest error:

You have bazel 0.11.1- installed.
./configure: line 669: /usr/local/cuda/extras/demo_suite/deviceQuery: No such file or directory
Configuration finished
Loading:
Loading: 0 packages loaded
Loading: 0 packages loaded
ERROR: /root/.cache/bazel/_bazel_root/05691b69ea8e69ef9e86d28a91a6ad71/external/io_bazel_rules_closure/closure/filegroup_external.bzl:23:16: The function 'set' has been removed in favor of 'depset', please use the latter. You can temporarily refer to the old 'set' constructor from unexecuted code by using --incompatible_disallow_uncalled_set_constructor=false
ERROR: /root/.cache/bazel/_bazel_root/05691b69ea8e69ef9e86d28a91a6ad71/external/io_bazel_rules_closure/closure/compiler/closure_js_binary.bzl:216:13: The function 'set' has been removed in favor of 'depset', please use the latter. You can temporarily refer to the old 'set' constructor from unexecuted code by using --incompatible_disallow_uncalled_set_constructor=false
ERROR: /root/.cache/bazel/_bazel_root/05691b69ea8e69ef9e86d28a91a6ad71/external/io_bazel_rules_closure/closure/compiler/closure_js_library.bzl:138:26: The function 'set' has been removed in favor of 'depset', please use the latter. You can temporarily refer to the old 'set' constructor from unexecuted code by using --incompatible_disallow_uncalled_set_constructor=false
ERROR: /root/.cache/bazel/_bazel_root/05691b69ea8e69ef9e86d28a91a6ad71/external/io_bazel_rules_closure/closure/compiler/closure_js_deps.bzl:44:13: The function 'set' has been removed in favor of 'depset', please use the latter. You can temporarily refer to the old 'set' constructor from unexecuted code by using --incompatible_disallow_uncalled_set_constructor=false
ERROR: error loading package '': Extension 'closure/compiler/closure_js_deps.bzl' has errors
ERROR: error loading package '': Extension 'closure/compiler/closure_js_deps.bzl' has errors
INFO: Elapsed time: 2.544s
FAILED: Build did NOT complete successfully (0 packages loaded)
The command '/bin/sh -c export PYTHON_BIN_PATH=`which python` &&     export PYTHON_LIB_PATH=/usr/lib/python2.7/dist-packages &&     export TF_NEED_MKL=0 &&     export CC_OPT_FLAGS="-march=native" &&     export TF_NEED_JEMALLOC=1 &&     export TF_NEED_GCP=0 &&     export TF_NEED_HDFS=0 &&     export TF_ENABLE_XLA=0 &&     export TF_NEED_OPENCL=0 &&     export TF_NEED_CUDA=1 &&     export TF_CUDA_CLANG=0 &&     export TF_CUDA_VERSION="7.5" &&     export CUDA_TOOLKIT_PATH="/usr/local/cuda" &&     export TF_CUDNN_VERSION="6" &&     export CUDNN_INSTALL_PATH=$CUDA_TOOLKIT_PATH &&     export TF_CUDA_COMPUTE_CAPABILITIES="3.5" &&     export TF_NEED_VERBS=0 &&     export TF_NEED_S3=0 &&     export TF_NEED_GDR=0 &&     export TF_NEED_OPENCL_SYCL=0 &&     export GCC_HOST_COMPILER_PATH=/usr/bin/gcc-4.9 &&     export TF_NEED_MPI=0 &&     export LD_LIBRARY_PATH=/usr/local/cuda/lib64/stubs:$LD_LIBRARY_PATH &&     ln -s /usr/local/cuda/lib64/stubs/libcuda.so /usr/local/cuda/lib64/stubs/libcuda.so.1 &&     cd / &&     wget https://github.com/tensorflow/tensorflow/archive/v1.3.1.tar.gz &&     tar xf v1.3.1.tar.gz &&     cd tensorflow-1.3.1 &&     ./configure &&     bazel build --local_resources 2048,2.0,1.0 -c opt --copt=-mavx --copt=-msse4.1 --copt=-msse4.2 --config=cuda tensorflow/tools/pip_package:build_pip_package &&     bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg &&     pip2 install /tmp/tensorflow_pkg/tensorflow-*.whl &&     cd / &&     rm -rf /v1.3.1.tar.gz &&     rm -rf /tensorflow-1.3.1 &&     rm -rf /tmp/tensorflow_pkg' returned a non-zero code: 1
Build Failed
[INFO] Finished streaming build
[ERROR] Container Builder exception encountered: Build failed with exit code 1

Not sure what to make of this.

[AdamSimpson]

I have fixed the issue by manually installing an appropriate JDK. As this is a mirror of an internal repository it will take an hour or two to sync up.