copydb

CopyDB

CopyDB is in-memory database works over Redis pub/sub approach. All writes are synced with Redis and then replicated across all CopyDB instances.

CopyDB is an embedded database and it gives you an ability to organize local cache inside your application.

CopyDB provides eventual consistency for data that updates frequently.

How it works

CopyDB downloads all available items from Redis after start. Then the database subscribes on items updates.

Each item contains a version to guarantee a consistent update on a particular instance. An item will be fully reloaded from Redis if a new version is not greater than existed by 1.

The database supports simple operations: SET, UNSET and REMOVE. It's described by Item interface:

type Item interface {
	Set(name string, data []byte)
	Unset(name string)
	Remove()
}

Read queries should be performed using embedded Query* functions:

• QueryByID
• QueryAll
• QueryPool
• QueryStats

Queries execution blocks items updates. Take the note you must write optimized queries.

Monitoring

CopyDB exposes essential metrics to get health status.

type Stats struct {
	ItemsApplied           int
	ItemsFailed            int
	ItemsEvicted           int
	ItemsReplicated        int
	VersionConfictDetected int
	DBScanned              int
}

Example

Please, checkout an example provided with the repository to figure out how to run your first application based on CopyDB.