Cloud Dataproc Initialization Actions

When creating a Dataproc cluster, you can specify initialization actions in executables and/or scripts that Dataproc will run on all nodes in your Dataproc cluster immediately after the cluster is set up. Initialization actions often set up job dependencies, such as installing Python packages, so that jobs can be submitted to the cluster without having to install dependencies when the jobs are run.

How initialization actions are used

Initialization actions must be stored in a Cloud Storage bucket and can be passed as a parameter to the gcloud command or the clusters.create API when creating a Dataproc cluster. For example, to specify an initialization action when creating a cluster with the gcloud command, you can run:

gcloud dataproc clusters create <CLUSTER_NAME> \
    [--initialization-actions [GCS_URI,...]] \
    [--initialization-action-timeout TIMEOUT]

During development, you can create a Dataproc cluster using Dataproc-provided regional initialization actions buckets (for example goog-dataproc-initialization-actions-us-east1):

REGION=<region>
CLUSTER=<cluster_name>
gcloud dataproc clusters create ${CLUSTER} \
    --region ${REGION} \
    --initialization-actions gs://goog-dataproc-initialization-actions-${REGION}/presto/presto.sh

⚠️ NOTICE: For production usage, before creating clusters, it is strongly recommended that you copy initialization actions to your own Cloud Storage bucket to guarantee consistent use of the same initialization action code across all Dataproc cluster nodes and to prevent unintended upgrades from upstream in the cluster:

BUCKET=<your_init_actions_bucket>
CLUSTER=<cluster_name>
gsutil cp presto/presto.sh gs://${BUCKET}/
gcloud dataproc clusters create ${CLUSTER} --initialization-actions gs://${BUCKET}/presto.sh

You can decide when to sync your copy of the initialization action with any changes to the initialization action that occur in the GitHub repository. Doing this is also useful if you want to modify initialization actions to meet your needs.

Why these samples are provided

These samples are provided to show how various packages and components can be installed on Dataproc clusters. You should understand how these samples work before running them on your clusters. The initialization actions provided in this repository are provided without support and you use them at your own risk.

Actions provided

This repository currently offers the following actions for use with Dataproc clusters.

• Install additional Apache Hadoop ecosystem components
  • Alluxio
  • Apache Drill
  • Apache Flink
  • Apache Gobblin
  • Apache Hive HCatalog
  • Apache Kafka
  • Apache Livy
  • Apache Oozie
  • Apache ZooKeeper
  • Presto
• Improve data science and interactive experiences
  • Miniconda
  • Apache Zeppelin
  • RStudio Server
  • Intel BigDL
  • Hue
• Configure the environment
  • Configure a nice shell environment
  • To switch to Python 3, use the conda initialization action
• Connect to Google Cloud Platform services
  • Install alternate versions of the Cloud Storage and BigQuery connectors. Specific versions of these connectors come pre-installed on Cloud Dataproc clusters.
  • Share a Cloud SQL Hive Metastore, or simply read/write data from Cloud SQL.
• Set up monitoring
  • Stackdriver
  • Ganglia

Removed actions

Previously, this repo provided init actions for the following, which have since been removed because equivalent functionality is now provided directly by Dataproc:

• Apache Tez: This is now pre-installed in all current Dataproc image versions.
• Datalab: Datalab has been replaced by Vertex AI Workbench, which integrates with Dataproc.
• Jupyter: This has been replaced with the Jupyter Optional Component.

Initialization actions on single node clusters

Single Node clusters have dataproc-role set to Master and dataproc-worker-count set to 0. Most of the initialization actions in this repository should work out of the box because they run only on the master. Examples include notebooks, such as Apache Zeppelin, and libraries, such as Apache Tez. Actions that run on all nodes of the cluster, such as cloud-sql-proxy, also work out of the box.

Some initialization actions are known not to work on Single Node clusters. All of these expect to have daemons on multiple nodes.

• Apache Drill
• Apache Flink
• Apache Kafka
• Apache Zookeeper

Feel free to send pull requests or file issues if you have a good use case for running one of these actions on a Single Node cluster.

Using cluster metadata

Dataproc sets special metadata values for the instances that run in your cluster. You can use these values to customize the behavior of initialization actions, for example:

ROLE=$(/usr/share/google/get_metadata_value attributes/dataproc-role)
if [[ "${ROLE}" == 'Master' ]]; then
  ... master specific actions ...
else
  ... worker specific actions ...
fi

You can also use the ‑‑metadata flag of the gcloud dataproc clusters create command to provide your own custom metadata:

gcloud dataproc clusters create cluster-name \
    --initialization-actions ... \
    --metadata name1=value1,name2=value2,... \
    ... other flags ...

For more information

For more information, review the Dataproc documentation. You can also pose questions to the Stack Overflow community with the tag google-cloud-dataproc. See our other Google Cloud Platform github repos for sample applications and scaffolding for other frameworks and use cases.

Mailing list

Subscribe to [email protected] for announcements and discussion.

Contributing changes

• See CONTRIBUTING.md

Licensing

• See LICENSE