google-cloud-cpp

HOWTO: using the Google Cloud Storage (GCS) C++ client in your project

This directory contains small examples showing how to use the GCS C++ client library in your own project. These instructions assume that you have some experience as a C++ developer and that you have a working C++ toolchain (compiler, linker, etc.) installed on your platform.

Before you begin

To run the quickstart examples you will need a working Google Cloud Platform (GCP) project and an existing bucket. The GCS quickstarts cover the necessary steps in detail. Make a note of the GCP project id and the bucket name as you will need them below.

Configuring authentication for the C++ Client Library

Like most Google Cloud Platform services, GCS requires that your application authenticates with the service before accessing any data. If you are not familiar with GCP authentication please take this opportunity to review the Authentication Overview. This library uses the GOOGLE_APPLICATION_CREDENTIALS environment variable to find the credentials file. For example:

Shell Command
Bash/zsh/ksh/etc. export GOOGLE_APPLICATION_CREDENTIALS=[PATH]
sh GOOGLE_APPLICATION_CREDENTIALS=[PATH];
export GOOGLE_APPLICATION_CREDENTIALS
csh/tsch setenv GOOGLE_APPLICATION_CREDENTIALS [PATH]
Windows Powershell $env:GOOGLE_APPLICATION_CREDENTIALS=[PATH]
Windows cmd.exe set GOOGLE_APPLICATION_CREDENTIALS=[PATH]

Setting this environment variable is the recommended way to configure the authentication preferences, though if the environment variable is not set, the library searches for a credentials file in the same location as the Cloud SDK. For more information about Application Default Credentials, see https://cloud.google.com/docs/authentication/production

Using with Bazel

  1. Install Bazel using the instructions from the bazel.build website.

  2. Compile this example using Bazel:

    cd $HOME/google-cloud-cpp/google/cloud/storage/quickstart
    bazel build ...
    

    Note that Bazel automatically downloads and compiles all dependencies of the project. As it is often the case with C++ libraries, compiling these dependencies may take several minutes.

  3. Run the example, change the placeholder to appropriate values:

    bazel run :quickstart -- [BUCKET NAME]
    

    Please use the plain bucket name. Do not include any gs:// prefix, and keep in mind the bucket naming restrictions. For example, bucket names cannot include forward slashes (/) or uppercase letters.

Using with CMake

  1. Install CMake. The package managers for most Linux distributions include a package for CMake. Likewise, you can install CMake on Windows using a package manager such as chocolatey, and on macOS using homebrew. You can also obtain the software directly from the cmake.org site.

  2. Install the dependencies with your favorite tools. As an example, if you use vcpkg:

    cd $HOME/vcpkg
    ./vcpkg install google-cloud-cpp[core,storage]
    

    Note that, as it is often the case with C++ libraries, compiling these dependencies may take several minutes.

  3. Configure CMake, if necessary, configure the directory where you installed the dependencies:

    cd $HOME/google-cloud-cpp/google/cloud/storage/quickstart
    cmake -S . -B .build -DCMAKE_TOOLCHAIN_FILE=$HOME/vcpkg/scripts/buildsystems/vcpkg.cmake
    cmake --build .build
    
  4. Run the example, changing the placeholder(s) to appropriate values:

    .build/quickstart [BUCKET NAME]
    

    Please use the plain bucket name. Do not include any gs:// prefix, and keep in mind the bucket naming restrictions. For example, bucket names cannot include forward slashes (/) or uppercase letters.

The gRPC plugin

The Google Cloud Storage client library includes a plugin to use gRPC as the transport to access GCS. For the most part, only applications with very large workloads (several Tbits/s of upload and/or download bandwidth) benefit from GCS+gRPC.

To enable the GCS+gRPC plugin you need to (a) link your application with an additional library, and (b) use a different function to initialize the google::cloud::storage::Client object. Both changes are illustrated in the quickstart_grpc program included in this directory.

Using with CMake

The GCS+gRPC plugin is disabled in CMake builds, this minimized the impact in build times and dependency management for customers not wanting to use the feature. To enable the feature add -DGOOGLE_CLOUD_CPP_STORAGE_ENABLE_GRPC=ON to your CMake configuration step. Using the previous example:

cd $HOME/google-cloud-cpp/google/cloud/storage/quickstart
cmake -S . -B .build -DGOOGLE_CLOUD_CPP_STORAGE_ENABLE_GRPC=ON \
    -DCMAKE_TOOLCHAIN_FILE=$HOME/vcpkg/scripts/buildsystems/vcpkg.cmake
cmake --build .build --target quickstart_grpc

Then run the additional program:

.build/quickstart_grpc [BUCKET NAME]

Using with Bazel

Run the example, change the placeholder to appropriate values:

bazel run :quickstart_grpc -- [BUCKET NAME]

Platform Specific Notes

macOS

gRPC requires an environment variable to configure the trust store for SSL certificates, you can download and configure this using:

curl -Lo roots.pem https://pki.google.com/roots.pem
export GRPC_DEFAULT_SSL_ROOTS_FILE_PATH="$PWD/roots.pem"

Windows

Bazel tends to create very long file names and paths. You may need to use a short directory to store the build output, such as c:\b, and instruct Bazel to use it via:

bazel --output_user_root=c:\b build ...

gRPC requires an environment variable to configure the trust store for SSL certificates, you can download and configure this using:

@powershell -NoProfile -ExecutionPolicy unrestricted -Command ^
    (new-object System.Net.WebClient).Downloadfile( ^
        'https://pki.google.com/roots.pem', 'roots.pem')
set GRPC_DEFAULT_SSL_ROOTS_FILE_PATH=%cd%\roots.pem