seaport-cache

A caching layer for seaport that lets you run nicely even if the seaport server dies.

This library takes a connected seaport object and caches all the results so that if the connection with seaport is lost, we can still run using the cached data.

Also adds some nice functions to automatically select a host based on random or round-robin principles.

example

Run the seaport service and register a server as normal.

On the client use like this:

var request = require('request');
var seaport = require('seaport');
var spcache = require('seaport-cache');

var ports = seaport.connect(9090, { secret : 'beep boop' });
var cache = spcache.getCache(ports);

setInterval(function () {
  cache.getRoundRobin('web@1.2.x', function (h) {
    var u = 'http://' + h.host + ':' + h.port;

    request(u).pipe(process.stdout);
  });
}, 1000);

methods

var seaportcache = require('seaport-cache');

seaportcache.getCache(ports)

Takes the ports object returned from seaport.connect and returns the cache object which you can query directly.

cache.get(role, cb)

Request an array of host/port objects through cb(services) that fulfill role

If there are no services that match the callback will be queued until some fulfilling role gets allocated.

cache.getFirst(role, cb)

As with get, but returns just the first host/port object instead of the whole array, equive to using services[0] in the callback.

cache.getAny(role, cb)

As with get but returns a random host/port object from the array.

cache.getRoundRobin(role, cb)

As with get but returns the first entry from the array on the first call, the second entry from the array on the second call, etc.

The round robin is based on the exact role string used, so different role queries get different sequences even if they match the name sequence of servers.