Cloudbreak for AWS and specifying RDS for HiveMetastore

[#1]

If using Cloudbreak to launch a cluster on AWS, can you specify a database (mysql, postgresql) in Amazon RDS to store the metadata for components such as Hive, Oozie, Ambari? What If I want to use RDS instead of having cloudbreak setup the databases locally on the cluster. Or is the option then to have Cloudbrek create the database locally on my cluster, and then change the config files to point to a new database in Amazon RDS (I will copy over metadata from existing local repo to RDS).

[#2 · Accepted solution]

Hi @Binu Mathew,

We have two different kind of external database support in Cloudbreak for provisioned cluster. One for Ambari (technical preview) and the other for Hive metastore, but unfortunately to use both of them you have to use Cloudbreak shell because the first one is officially available only in Cloudbreak and the second is HDCloud exclusive feature.

To configure RDS for Ambari you have to execute database configure shell command before executing cluster create command:

database configure --vendor --host --port --name --username --password

Hive metastore RDS configuration available in cluster create command:

cluster create --databaseType --connectionUserName --connectionPassword --connectionURL --hdpVersion

[update]

RDS configuration for Cloudbreak is configurable in Cloudbreak deployer's Profile file by adding the environment variables below:

CB_DB_PORT_5432_TCP_ADDR
CB_DB_PORT_5432_TCP_PORT
CB_DB_ENV_USER
CB_DB_ENV_PASS
CB_DB_ENV_DB