Cloud Clustering for Datastores
blakebyrnes opened this issue · 0 comments
blakebyrnes commented
These are my notes for enabling clustering for the Cloud module:
- #184
- no more db file in the package, just the endpoint
- storageEngineEndpoint can go directly into datastore code
- no more db file in the package, just the endpoint
- Datastore Migrations
- Default migration copies previous db and creates a new one
- Handle multiple versions in Ulixee.app
- Database Service:
- Exposes an ability to query an engine with pre-fetched input/outputs for tables/virtual tables
- Db Creation api (migrate, copy, etc)
- Only for "internal" sqlite queries
- Can require payment
- Public api to ask for storageEngineEndpoint
- Nodes point to "lead". Lead returns an api call that returns SaaS service endpoints:
- storage service, db service, node lookup service
- lead subscribes (optionally) to the broader kad
- Datastore lookup function to find a datastore version in a cluster
- Use the cluster's SaaS storage service
- Server will cache locally