Registry: Read-only backup server

[jpmckinney]

The production server is capable of collecting and processing data from its data sources. The output of these processes are a small data_registry database in PostgreSQL (MBs, not GBs), and json.gz files for monthly, yearly, and full dumps of the data in /data/storage/exporter_dumps (currently tens of GBs).

The backup server, on the other hand, does not need to be capable of collecting and processing data. It just needs a copy of the data_registry database and copies of the json.gz files – so that the data registry application can allow users to download the JSON files. The application has very low CPU and RAM requirements.

---

It also hosts a flattening tool (spoonbill), which is an online wizard to transform the JSON to Excel/CSV. Having the JSON files on the same server make this second tool much faster. Users can also upload their own JSON files to the flattening tool.

For the flattening tool, its processing is fairly CPU intensive, so it needs to be comparable to the production server. The requirements expressed by Quintagroup were "4 or more CPU core" "RAM - 8GB or more".

Users of the flattening tool can download the Excel/CSV output for a given time period. If we want to preserve access, we will need to copy over some files as well. I'll also need to check if there's a database to sync.