tunneler: Connect to databases via ssh tunnel

A command tool and library to connect to databases with minimal fuss.

To install:

$ pip install tunneler

You will probably also need the database drivers for your preferred database. To install the postgres ones automatically when you install tunneler, do this:

$ pip install tunneler[pg]

To run:

$ tunneler name_of_task name_of_connection

As the above command shows, there are two concepts to understand when running tunneler, connections and tasks.

Define connections in a connections.yaml file in your current directory, as follows:

prod:
  remote_host: publicjumpboxservername
  ssh_username: username
  ssh_pkey: |
    -----BEGIN RSA PRIVATE KEY-----
    .......................................
    .......................................
    -----END RSA PRIVATE KEY-----
  private_dburl: postgres://username:password@privatedbservername/databasename
  local_port: 5433

dev:
  .....

You can also specify an ssh_config variable in the config. tunneler will then use the values from that hosts ssh config. An example:

prod:
  ssh_config: prod-jumpbox
  private_dburl: postgres://username:password@privatedbservername/databasename
  local_port: 5433

dev:
  .....

This will use the remote host/username/private key specified in your existing ssh config so you don't need to repeat them in the config file

Tunneler will use this information to set up a local tunneled post where you can access this database directly.

A connection URL to this local port will then be provided to each task.

There are three inbuilt tasks:

• test_connection (run this one to check your config is correct)
• do_full_pg_dump (assumes you have pg_dump installed)
• do_schema_pg_dump (also assumes you have pg_dump installed)

So for example to dump the production schema we'd run the following command:

$ tunneler do_schema_pg_dump prod

Custom tasks

Want to define your own tasks? Simply create a tasks.py file (or module) in the current directory.

Define a top level method in this file that accepts a connection url, and that method name will be available as a task.

For instance, the test connection task could be re-implemented as follows:

from sqlbag import S

def test_connection_custom(dburl):
    with S(dburl) as s:
        s.execute('select 1')

You'd then run this as follows:

$ tunneler test_connection_custom prod

Pretty simple.