Automatically discover your nodejs instances with built-in support for automatic single master and capability advertising.
Automatic and decentralized discovery and monitoring of nodejs instances with built in support for a variable number of master processes, service advertising and channel messaging.
So, you have a whole bunch of node processes running but you have no way within each process to determine where the other processes are or what they can do. This module aims to make discovery of new processes as simple as possible. Additionally, what if you want one process to be in charge of a cluster of processes? This module also has automatic master process selection.
This module uses broadcast and multicast features from node's dgram module. All required features of the dgram module are implemented in the following versions of node
- v0.4.x - v0.6.9+
Be sure to look in the examples folder, especially at the distributed event emitter
var Discover = require'node-discover';var d = Discover;don"promotion"/** Launch things this master process should do.** For example:* - Monitior your redis servers and handle failover by issuing slaveof* commands then notify other node instances to use the new master* - Make sure there are a certain number of nodes in the cluster and* launch new ones if there are not enough* - whatever**/console.log"I was promoted to a master.";;don"demotion"/** End all master specific functions or whatever you might like.**/console.log"I was demoted from being a master.";;don"added"console.log"A new node has been added.";;don"removed"console.log"A node has been removed.";;don"master"/** A new master process has been selected** Things we might want to do:* - Review what the new master is advertising use its services* - Kill all connections to the old master*/console.log"A new master is in control";;
npm install node-discover
git clone git://github.com/wankdanker/node-discover.git
DiscoverhelloInterval : How often to broadcast a hello packet in milliseconds; Default: 1000checkInterval : How often to to check for missing nodes in milliseconds; Default: 2000nodeTimeout : Consider a node dead if not seen in this many milliseconds; Default: 2000masterTimeout : Consider a master node dead if not seen in this many milliseconds; Default: 2000address : Address to bind to; Default: '0.0.0.0'port : Port on which to bind and communicate with other node-discover processes; Default: 12345broadcast : Broadcast address if using broadcast; Default: '255.255.255.255'multicast : Multicast address if using multicast; Default: null don't use multicast, use broadcast)mulitcastTTL : Multicast TTL for when using multicast; Default: 1unicast : Comma separated String or Array of Unicast addresses of known nodes. It is advised to specify the `address` of the local interface when using unicast and expecting local discovery to workkey : Encryption key if your broadcast packets should be encrypted; Default: null (that means no encryption);mastersRequired : The count of master processes that should always be available,weight : A number used to determine the preference for a specific process to become master. Higher numbers win. Default : Discover.weight()client : When true operate in client only mode (don't broadcast existence of node just listen and discover; Default : falsereuseAddr : Allow multiple processes on the same host to bind to the same address and port Only applies to node v012+ Default: trueignoreProcess : If set to false will not ignore messages from other Discover instances within the same process on non-reserved channels join will receive them Default: trueignoreInstance : If set to false will not ignore messages from self on non-reserved channels join will receive them Default: truecallback;
Promote the instance to master.
This causes the old master to demote.
var Discover = require'node-discover';var d = Discover;dpromote;
Demote the instance from being a master. Optionally pass true to demote to specify that this node should not automatically become master again.
This causes another node to become master
var Discover = require'node-discover';var d = Discover;ddemote; //this node is still eligible to become a master node.//orddemotetrue; //this node is no longer eligible to become a master node.
Join a channel on which to receive messages/objects
var Discover = require'node-discover';var d = Discover;//Pass the channel and the callback function for handling received data from that channelvar success = djoin"config-updates"if dataredisMaster//connect to the new redis master;if !success//could not join that channel; probably because it is reserved
Leave a channel
var Discover = require'node-discover';var d = Discover;//Pass the channel which we want to leavevar success = dleave"config-updates";if !success//could leave channel; who cares?
Send a message/object on a specific channel
var Discover = require'node-discover';var d = Discover;var success = dsend"config-updates" redisMaster : "10.0.1.4" ;if !succes//could not send on that channel; probably because it is reserved
Advertise an object or message with each hello packet; this is completely arbitrary. Make this object/message whatever applies to your application that you want your nodes to know about the other nodes.
var Discover = require'node-discover';var d = Discover;dadvertiselocalServices :type : 'http' port : '9911' description : 'my awesome http server'type : 'smtp' port : '25' description : 'smtp server';//ordadvertise"i love nodejs";//ordadvertise something : "something" ;
Start broadcasting hello packets and checking for missing nodes (start is called automatically in the constructor)
var Discover = require'node-discover';var d = Discover;dstart;
Stop broadcasting hello packets and checking for missing nodes
var Discover = require'node-discover';var d = Discover;dstop;
For each node execute fn, passing fn the node fn(node)
var Discover = require'node-discover';var d = Discover;deachNodeif nodeadvertisement == "i love nodejs"console.log"nodejs loves this node too";;
Each event is passed the
Node Object for which the event is occuring.
Triggered when the node has been promoted to a master.
Triggered when the node is no longer a master.
Triggered when a new node is discovered
Triggered when a new node is not heard from within
Triggered when a new master has been selected
isMaster: trueisMasterEligible: trueadvertisement: nulllastSeen: 1317323922551address: '10.0.0.1'port: 12345id: '31d39c91d4dfd7cdaa56738de8240bc4'hostName : 'myMachine'
I have not tested large packets. The current version does not handle recombining split messages.
Copyright (c) 2011 Dan VerWeire email@example.com
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.