Requestador

Demo

See this youtube video for an example implenetation in drupal. This is only a fraction of what this software does. https://www.youtube.com/watch?v=4fv0t62wPf0

Intro

Sometimes you need POSTs to be GET's, POST to become socket messages or just listen on an amqp queue an put that stuff out there (POST/GET/socket)

This is what Requestador does for you.

You could call it a relay, or a hub a messagebroker or whatever.

HTTP / SOCKET / (rabbitMQ) amqp mediator

Installation

• npm install
• configure settings in data/config.ini and data/amqp.ini
• admin login happens via PAM
• create data/listeners.inc
• create data/senders.inc

Requirements

• are in package.json.
• I run it on node v0.10.33, so anything more recent should work.

Configuration

rename your amqp.ini.example to amqp.ini and add your servers. Section of AMQP settings

[amqpserver1]
name=amqp1
ip=127.0.0.1
port=5672
user=guest
password=guest
heartbeat=25
vhost=/
[amqpserver2]
name=amqp2
ip=127.0.0.1
port=5672
user=guest
password=guest
heartbeat=25
vhost=/

rename your config.ini.example to config.ini and set your settings.

Section of amqp settings

[amqp]
useamq=true

Section of socket settings

[server]
usesocketio=true
ip=127.0.0.1
port=3000

Section of Webserver settings

[web]
useweb=true
ip=127.0.0.1
port=3000

Section of Proxy settings Enable routing all outgoing requests through a proxy server.

[proxy]
proxy_enabled=true
proxy_server=127.0.0.1
proxy_port=3000

Section of Logging settings

[log]
logFile=/var/log/requestador.log
logToFile=true
logToConsole=true
failedRequests=/var/log/failedRequests.log

Section of Monitoring settings

[monitor]
logFile=/var/log/requestador.log
active=true
sendmail=true
polltimeout=3000 (milliseconds)
monitorMailFrom[email protected]
monitorMailTo[email protected]
monitorMailSubject=Er was een probleem
sendmailPath=/usr/sbin/sendmail
sendmailArgs= -t -f [email protected]

Section of Autostart configuration

[autostart]
smtphost=127.0.0.1
smtpport=25
ignore_tls=true
min_sleeptime=1000
min_uptime=10000

Sleeptime is the time slept between restarting the process (defaults to 1000) min_uptime is the time the process is required to be up to be able to restart. (defaults to 1000) if the proces can not be up 1000 ms, it is problematic.

Run

If you want it running continuously (forever), and restart after craches

• node server_index.js

This is better than but uses the same logic as

• forever -c node server_index.js

connects to amqp , creates webserver, and socket server.

Listeners

Listeners are configured in listeners.inc The formatting of this file is like this:

name | method | uri | senders_csv

• name: a unique identifier for a listener
• method: POST/GET/SOCKET/AMQP
• uri: The socket channel "testchannel" or a certain uri "/testuri" on the current domain.
  • foramqp this is in the format "amqpserver:QUEUE:key:exchange"
  • e.g. amqp1:/queue/someQueueName:#.be.test.key:requestador.topic
• senders: a comma separated list of SENDER NAMES eg(DOPOST,DOSOCKET,DOAMQP).

Senders

Senders are configured in senders.inc The formatting of this file is like this:

name | method | url | data

• name: A unique name for this sender
• method: POST/GET/SOCKET/AMQP
• url: the full url to POST/GET to or the channel to SOCKET/AMQ to e.g. "tstchannel" or "http://mydomain/url/to_/post/to"
• data: Allowerd datatypes: forward (forward postbody or socket msg) empty or fixed

Monitoring

It's best to connect Nagios to the polling path of the server: "/requestadorPoll" If you dont have that, you can run "node monitor.js" This script will alert you via email if the servers stops running. It's best to configure (and test) your monitoring in config.ini before you use it in production.