Principles
seneca-archive plugs into your seneca data flow, intercepts archived objects and acts as if it is the principal database.
Archiving requires 2 seneca instances. That allows to have separate configurations and flows for these 2 instances. The principal instance is the one that your use for regular DB transactions. seneca-archive is layered into this principal instance.
When a request is unsatisfied in the principal instance, seneca-archive attempts to run the query against the secondary instance. If unsuccessful, seneca-archive will return the result from the principal instance.
For your application, handling archived entities is completely transparent.
Performance considerations
- For regular objects, the process has very little overhead.
- For 'misses' the process has maximum overhead of hitting 2 entity chains. A miss is a request on an entity that exists in neither the principal nor the secondary database.
- For an archived entity, the overhead is to first attempt to access the entity in the principal DB and then the secondary one.
- When searching multiple entities, queries are run against all DBs. This negative overhead can be cancelled by the use of a 3rd party indexing engine (eg. elasticsearch)
Installation
npm i --save seneca-archive
You will also need seneca.
Usage
var oneMonthAgo = ...
var oneYearAgo = ...
seneca.use('archive', {
archivalInstance: senecaInstance,
directArchival: true,
checkExistsBeforeUpdate: true,
conditions: {
'sys/user': [
{ lastLogin: { $lt: oneYearAgo } },
{ lastLogin: null, active: false, registrationDate: oneMonthAgo }
]
}
})
Options
archivalInstance
This is the seneca secondary instance used to manage the archived objects.
directArchival
if directArchival is enabled/disabled (enabled by default). When anabled, new entities are checked against the archival conditions. If they match, these entities are directly saved in the archival DB.
checkExistsBeforeUpdate
check in which DB the updated entity lives. true by default. This should be set to false if the DB driver throws an error when a non existing entity is updated.
conditions
For each entity type that should be archived, a list of conditions that should be met for an entity to be archived.
Triggering the archival process
The archival process should be triggered externally. seneca-archive does not have it's own cron like schedule so that's up to you.
Archiving consists in doing a full database scan for each entity and then moving the right entities into the secondary DB.
seneca.act({role: 'archive', cmd: 'scan', skip: 0, limit: 100, entity: 'sys/user'}, function(args, done) {
})
The option entity
is optional. If not set, the archival check will apply to all entity types defined in the plugin's global
options.