Configuring Alluxio with Google Cloud Storage

This guide describes how to configure Alluxio with Google Cloud Storage (GCS) as the under storage system.

Initial Setup

First, the Alluxio binaries must be on your machine. You can either compile Alluxio, or download the binaries locally.

Then, if you haven’t already done so, create your configuration file with bootstrapConf command. For example, if you are running Alluxio on your local machine, ALLUXIO_MASTER_HOSTNAME should be set to localhost

$ ./bin/alluxio bootstrapConf <ALLUXIO_MASTER_HOSTNAME> gcs

Alternatively, you can also create the configuration file from the template and set the contents manually.

$ cp conf/alluxio-env.sh.template conf/alluxio-env.sh

Also, in preparation for using GCS with Alluxio, create a bucket (or use an existing bucket). You should also note the directory you want to use in that bucket, either by creating a new directory in the bucket, or using an existing one. For the purposes of this guide, the GCS bucket name is called GCS_BUCKET, and the directory in that bucket is called GCS_DIRECTORY.

If you are new to GCS, please read its documentations.

Configuring Alluxio

You need to configure Alluxio to use GCS as its under storage system. The first modification is to specify an existing GCS bucket and directory as the under storage system by modifying conf/alluxio-site.properties to include:

alluxio.underfs.address=gs://GCS_BUCKET/GCS_DIRECTORY

Next, you need to specify the Google credentials for GCS access. In conf/alluxio-site.properties, add:

fs.gcs.accessKeyId=<GCS_ACCESS_KEY_ID>
fs.gcs.secretAccessKey=<GCS_SECRET_ACCESS_KEY>

Here, <GCS_ACCESS_KEY_ID> and <GCS_SECRET_ACCESS_KEY> should be replaced with your actual GCS interoperable storage access keys, or other environment variables that contain your credentials. Note: GCS interoperability is disabled by default. Please click on the Interoperability tab in GCS setting and enable this feature. Then click on Create a new key to get the Access Key and Secret pair.

Alternatively, these configuration settings can be set in the conf/alluxio-env.sh file. More details about setting configuration parameters can be found in Configuration Settings.

After these changes, Alluxio should be configured to work with GCS as its under storage system, and you can try Running Alluxio Locally with GCS.

Configuring Application Dependency

When building your application to use Alluxio, your application will have to include the alluxio-core-client module. If you are using maven, you can add the dependency to your application with:

<dependency>
  <groupId>org.alluxio</groupId>
  <artifactId>alluxio-core-client</artifactId>
  <version>1.4.0</version>
</dependency>

Configuring Distributed Applications Runtime

When I/O is delegated to Alluxio workers (i.e., Alluxio configuration alluxio.user.ufs.operation.delegation is true, which is false by default since Alluxio 1.1), you do not have to do any thing special for your applications. Otherwise, since you are using an Alluxio client that is running separately from the Alluxio Master and Workers (in a separate JVM), then you need to make sure that your Google credentials are provided to the application JVM processes as well. There are different ways to do this. The first approach is to add them as command line options when starting your client JVM process. For example:

$ java -Xmx3g -Dfs.gcs.accessKeyId=<GCS_ACCESS_KEY_ID> -Dfs.gcs.secretAccessKey=<GCS_SECRET_ACCESS_KEY> -cp my_application.jar com.MyApplicationClass myArgs

Alternatively, you may copy conf/alluxio-site.properties (having the properties setting credentials) to the classpath of your application runtime (e.g., $SPARK_CLASSPATH for Spark), or append the path of this site properties file to the classpath.

Running Alluxio Locally with GCS

After everything is configured, you can start up Alluxio locally to see that everything works.

$ ./bin/alluxio format
$ ./bin/alluxio-start.sh local

This should start an Alluxio master and an Alluxio worker. You can see the master UI at http://localhost:19999.

Next, you can run a simple example program:

$ ./bin/alluxio runTests

After this succeeds, you can visit your GCS directory GCS_BUCKET/GCS_DIRECTORY to verify the files and directories created by Alluxio exist. For this test, you should see files named like:

GCS_BUCKET/GCS_DIRECTORY/alluxio/data/default_tests_files/Basic_CACHE_THROUGH

To stop Alluxio, you can run:

$ ./bin/alluxio-stop.sh all

GCS Access Control

If Alluxio security is enabled, Alluxio enforces the access control inherited from underlying object storage.

The GCS credentials specified in Alluxio config represents a GCS user. GCS service backend checks the user permission to the bucket and the object for access control. If the given GCS user does not have the right access permission to the specified bucket, a permission denied error will be thrown. When Alluxio security is enabled, Alluxio loads the bucket ACL to Alluxio permission on the first time when the metadata is loaded to Alluxio namespace.

Mapping from GCS user to Alluxio file owner

By default, Alluxio tries to extract the GCS user id from the credentials. Optionally, alluxio.underfs.gcs.owner.id.to.username.mapping can be used to specify a preset gcs owner id to Alluxio username static mapping in the format “id1=user1;id2=user2”. The Google Cloud Storage IDs can be found at the console address. Please use the “Owners” one.

Mapping from GCS ACL to Alluxio permission

Alluxio checks the GCS bucket READ/WRITE ACL to determine the owner’s permission mode to a Alluxio file. For example, if the GCS user has read-only access to the underlying bucket, the mounted directory and files would have 0500 mode. If the GCS user has full access to the underlying bucket, the mounted directory and files would have 0700 mode.

Mount point sharing

If you want to share the GCS mount point with other users in Alluxio namespace, you can enable alluxio.underfs.object.store.mount.shared.publicly.

Permission change

In addition, chown/chgrp/chmod to Alluxio directories and files do NOT propagate to the underlying GCS buckets nor objects.

Need help? Ask a Question