search

matt-lebl / uvic-marketplace

5 stars 0 forks source link

SCA: Implement database replication #210

Open Broondoon opened 5 months ago

Broondoon commented 5 months ago

Plan and establish a worker/controller relationship between the multiple database instances. Coordinate with (hopefully) existing encryption / hashing procedure to so that write requests are sent to the controller db and read requests are sent to any of the worker db.

If encryption is not set up by this point, design this implementation with encryption possibilities in mind.

Broondoon commented 5 months ago

Further to SCA: Implement data sharding #209, moving this into Subsprint 3 for the same reasons listed there.

Broondoon commented 4 months ago

Writeahead log - no changes made to db filesystem before entire transaction written to a log. Some program listens to this log and sends to another instance of the db. Kubernetes would have to be configured to expose the specific dbs via labels (app selector thing).

Maybe leave this until we've found where the bottlenecks are. This is a lot of extra work and our db is as I'm told quite resilient already.