Merge pull request #335 from OperationCode/document-backups

Add instructions for manual DB backups

Author: irvingpop

Ops.md

@@ -18,3 +18,20 @@ open https://localhost:8443
   - Validate the staging environment (try a few API calls, particularly GET and POST)
 4. Repeat those steps for the production environment
   - When validating prod, please add a real resource to the API, not a test resource (there's no way to delete a test resource using the API)
+
+# Backing up prod to staging
+
+Every so often, we should back up prod to staging to ensure that we don't lose data, and also so that the staging data isn't stale during testing. And also, to blow away test changes that were made to staging while validating a new image version before going to prod.
+
+1. Create a postgres image in the cluster
+```bash
+kubectl run -it postgres --image postgres:9.6 --rm -- /bin/bash
+```
+2. Run the following command to dump the resources data (replace username and dbname with the values in resources-api-secrets). You'll be prompted for the db password, so have that value ready as well.
+```bash
+pg_dump --clean --if-exists --inserts --quote-all-identifiers --host=python-prod.czwauqf3tjaz.us-east-2.rds.amazonaws.com --username=XXXX --dbname=XXX -f prod_backup.sql
+```
+3. Restore the data to staging with this command (have the password for staging db at the ready, because again, you'll be prompted for it)
+```bash
+psql --host=python-staging.czwauqf3tjaz.us-east-2.rds.amazonaws.com --username=XXXX --dbname=XXXX -f prod_backup.sql
+```