PaymentGateway node manager [PG_nm]

PG_nm is stateless async microservice with Redis RPC interaction

Components

• node.js vertical cluster ([MASTER] fork [WORKER] processes by worler types)
• [MASTER] Redis RPC channel event handler. Cluster CMD routing
• [WORKER Checker] exec cluster RPC CMD.
• [WORKER Scheduled] task scheduler/runner/manager

Master process behavior

 @Master process behavior
 - init cluster workers
 - respawn workers on 'die'
 0. [MASTER] init RPC channel connection. Pass CMDs from RedisRPC to workers
 1. [MASTER] handel RPC channel events
 2. [MASTER] pass events to workers using routing
 3. [MASTER] Forward events from worker to worker using routing
 4. [WORKER] do something
 5. [MASTER] handle MSG from worker
 6. [MASTER] exec RPC callback done(err,data)

Scheduler process behavior

 @Scheduler (worker 1):
 1. handle RPC messages/CMDs from Master node
 2. execute CMDs
 [CMD executor]:
 - [cluster RPC CMD] `bootstrap` => create task list and apply tasks (cold start)
 - [cluster RPC CMD] `taskList`
 - [cluster RPC CMD] `taskStop`
 - [cluster RPC CMD] `taskRun`
 - [cluster RPC CMD] `taskAdd`
 [Task scheduler]:
 - addTask to task list
 - apply task
 - run task
 [current tasks]:
 1. check -> check all nodes by timeout (send cluster RPC CMD to worker @checker)

Checker process behavior

 @Checker (worker 2)
 1. handle RPC messages/CMDs from Master node
 2. execute CMDs
 [CMD executor]:
 - [cluster RPC CMD] `bootstrap` => node CFG bootstrapper (load node cfg from file to DB) (cold start)
 - [cluster RPC CMD] `check` => checkNode(type, cfg) -> exec RedisRPC to pg_jrpc-proxy -> result will update DB node status
 - [cluster RPC CMD] `getBestNode(type)` -> get best node from DB
 - [cluster RPC CMD] `getNodes` => get all nodes configs from DB
 - [cluster RPC CMD] `getNodeConfig` by ID/nodeHash from DB
 - [cluster RPC CMD] `addNode(type, config)` to DB
 - [cluster RPC CMD] `rmNode by ID/nodeHash` from DB
 - [cluster RPC CMD] `updateNode by ID/nodeHash` in DB

architecture