This initialization action installs the latest version of Apache Tez on a Google Cloud Dataproc cluster. In this initialization action:

1. Required libraries such as protobuf are installed
2. Tez is compiled
3. Tez is put into HDFS
4. Hadoop is configured to work with Tez on the Dataproc cluster

You can use this initialization action to create a new Dataproc cluster with Apache Tez installed by:

1. Uploading a copy of this initialization action (tez.sh) to Google Cloud Storage.

2. Using the gcloud command to create a new cluster with this initialization action. Note - you will need to increase the initialization-action-timeout to at least 15 minutes to allow for Tez and its dependencies to compile and install. The following command will create a new cluster named <CLUSTER_NAME>, specify the initialization action stored in <GCS_BUCKET>, and increase the timeout to 15 minutes.

   gcloud dataproc clusters create <CLUSTER_NAME> \
   --initialization-actions gs://<GCS_BUCKET>/tez.sh   
   --initialization-action-timeout 15m

3. Once the cluster is created, Tez should work with your Cloud Dataproc jobs.

You can find more information about using initialization actions with Dataproc in the Dataproc documentation.

• This script has a number of user-defined variables at the top, such as the Tez version and install location.
• This script builds from source which may cause issues if you use an unstable release.
• Be careful with the protobuf version - Tez often requires a very specific version to function properly.