Prerequisites

1. Have a technical resource person from CSSAT provide you with the private key for the bastion host for the dcyf-cssat collaboration VPC in AWS. Here, we assume that the file is named dcyf-cssat-ssh-key.pem. If you copy the key out of LastPass, there may be blank lines inserted between the "Begin" and "End" lines and the actual key--if so, you will need to delete them.

2. Have a technical resource person from CSSAT white list your IP address on the VPC.

3. Make sure you chmod the key to ensure that it isn't publicly viewable.

chmod 400 dcyf-cssat-ssh-key.pem

4. Have a technical resource person from CSSAT provide you with a username and password for the sprout_replica Postgres database. If you are looking to make changes to the curated schema, you will need to be logged in using the postgres user.

Accessing the database

Access via the bastion host

Connect to the bastion server.

ssh -i dcyf-cssat-ssh-key.pem [email protected]

Connect to the database using the psql client on the bastion server.

psql postgresql://<User>:<Password>@cssat-dcyf-sprout-replica.crartr7yq7ee.us-east-1.rds.amazonaws.com/sprout_replica

Access via port fowarding

Setup port forwarding

 ssh -i "dcyf-replica-user.pem" -f -N -L 5433:cssat-dcyf-sprout-replica.crartr7yq7ee.us-east-1.rds.amazonaws.com:5432 [email protected] -v

Connect to the database locally using your favorite Postgres client. For example, using a the psql terminal client, the following connection string will provide you with access to the replica.

psql postgresql://<User>:<Password>@localhost:5433/sprout_replica

Changing the database

1. Login to aptible and find the credentials for the production sprout database.

2. Add the aptible credentials to an .env file with variables specified as shown below.

CSSAT_DCYF_SPROUT_REPLICA_PWD=<postgres user password>
SPROUT_DB_NAME=<aptible database name>
SPROUT_DB_SERVER=<aptible database server>
SPROUT_DB_PORT=<aptible database port>
SPROUT_DB_USER=<aptible username>
SPROUT_DB_PWD=<aptible password>

3. If your new materialized view requires access to a new relation in the production sprout database, edit the schema import in sprout_fdw_config.sql.

NOTE: If you can write the new materialized view within the existing schema, skip this step and proceed to step 4.

4. Define the new materialized view (or update an existing view) in the curated schema.

5. Grant dcyf-replica-user access to the materialized view by updating dcyf_grants.sql.

NOTE: If you are just updating an existing materialized view, skip this step and proceed to step 8.

6. Schedule a nightly refresh of the materialized view by updating schedule_refresh.sql.

7. Update rebuild_views.sh to include a build of your materialized view when the script is run.

8. Rebuild the materialized views by running

cd scripts
./rebuild_views.sh

NOTE: Step 8 accomplishes two things: A. If your materialized view is new, it creates the view so that it can be picked up by the pg_cron that you established in Step 6, and B. It helps to make sure that your materialized view contains all of the necessary relation dependencies.