multiple-datasets.md

Working With Multiple Data Sets

You can deploy more than one source{d} CE instance with different sets of organizations, or repositories, to analyze.

For example, you may have initially started source{d} CE with the repositories in the src-d organization, with the command:

$ sourced init orgs --token <token> src-d

After a while, you may want to analyze the data on another set of repositories. You can run sourced init again with a different organization:

$ sourced init orgs --token <token> bblfsh

This command will then stop all the running containers used for the previous dataset, create an isolated environment for the new data, and create a new, clean deployment.

Please note that each path will have an isolated deployment. This means that for example any chart or dashboard created for the deployment for src-d will not be available to the new deployment for bblfsh.

Each isolated environment is persistent (unless you run sourced prune). Which means that if you decide to re-deploy source{d} CE using the original organization:

$ sourced init orgs --token <token> src-d

You will get back to the previous state, and things like charts and dashboards will be restored.

These isolated environments also allow you to deploy source{d} CE using a local set of Git repositories. For example, if we want a third deployment to analyze repositories already existing in the ~/repos directory, we just need to run init again:

$ sourced init local ~/repos

You can list all the installed instances, and know which one is active at any moment by running sourced workdirs.

If you are familiar with Docker Compose and you want more control over the underlying resources, you can explore the contents of your ~/.sourced directory. There you will find a docker-compose.yml and .env files for each set of repositories used by sourced init.

You can read more about how the environments are isolated in the source{d} CE architecture docs