The Cluster Manager uses a database to store the data associated with several important features: user profiles, API keys, job history, batch definitions, and batch data.
The Cluster Manager can be connected to two types of database servers:
- MongoDB version 4.0 or later, deployed on-premise, on the Cloud, or hosted by a SaaS provider.
- Amazon Web Services (AWS) DocumentDB 4.0 or later, when deployed to AWS.
MongoDB offers various configuration options, including clustering and encryption over the wire and at rest. Please follow the steps as explained in the official MongoDB documentation. The installation steps are well explained for each platform. When the installation is complete, the MongoDB daemon mongod should be running and ready for connections on port 27017 (the default). One other option, if you want to avoid installing the database yourself, would be to sign up for a hosted solution in the Cloud.
You can also set up an AWS DocumentDB 4.0 database if you are deploying the Cluster Manager on the AWS platform (please, refer to their developer guide). DocumentDB offers various configuration options, including clustering and encryption over the wire and at rest. Note that AWS DocumentDB cannot be installed on-premises.
Once MongoDB or DocumentDB is installed or provisioned in the Cloud, please make a note of the connection string (which is a URL that will be provided to you during installation). We will need this URL to connect the Cluster Manager to the database. The URL has the following form. You should of course substitute the MongoDB username and password that you chose when installing the database:
If you have installed MongoDB on your personal machine for testing purposes, the connection string URL should be the following:
You can check that your connection string is correct by using the mongo shell and providing the connection string as the first argument:
> mongo mongodb://localhost:27017 MongoDB shell version v4.0.4 connecting to: mongodb://localhost:27017 MongoDB server version: 4.0.4
As noted earlier, you need to install version 4.0 or later. If you try an earlier version, you may see connection error messages like the following when you try to start the Cluster Manager:
fatal : Failed to connect to database: server selection error: server selection timeout current topology: Type: Unknown