dropbox / amqp-coffee Goto Github PK

Hi, I have an issue where a blank message (0 size, empty body) published to a queue I'm consuming with amqp-coffee stops the consumer.
I'm using QOS mode to validate each message and then reject it or process and ack it. When a blank message gets published to the queue, my message handler is not fired and the message can be neither acked nor rejected – the consumer just stops working. When I disable prefetchCount, the empty messages are discarded silently.
My guess is that somewhere a callback is not called and everything just stops. I tried debugging it but couldn't get anywhere. Maybe it's something you can look into? I'll gladly help.

The uuid module replaces node-uuid (same author) and issues a warning upon installing node-uuid.

npm WARN deprecated [email protected]: use uuid module instead

Hey there.

I've been using the wrapper in prod for almost a week now, and for the most part, it's been working great.

I noticed today that every few hours, an unspecified error event is thrown from Connection.js. I've included the stack trace, and screenshots from my logs. Please, let me know if there's anything else I can provide...

Stack Trace Gist: https://gist.githubusercontent.com/schonfeld/271cf723bed2385aa02d/raw/65a18c13552ad0053482383751e44ede1185d683/gistfile1.txt

Screenshot: https://www.dropbox.com/s/gm5geyrf6pyp1mg/Screenshot%202014-08-11%2012.43.29.png

Thanks,

• Michael.

Hello, I finally got a chance to try this out, and I've got some small feedback and a few larger issues i've found.

I'll just note them here for now, I can write up with more detail in separate issues if you want

• Error handling: Socket errors after the initial connection are silently discarded
• Timeouts: There is currently no built-in way to time out a connection attempt
• Client Info: This is hard-coded, it would be good to allow callers to override these
• AMQP URI: It would be nice to support the amqp uri format for connecting: amqp://user:pass@host:port/vhost
• I can see why you would want to separate the Exchange object from the exchange.declare call, but as both take callbacks it ends up being a little fiddly. I would suggest that either the object creation has no side effects, or providing a convenience method for the common case of declaring an exchange
• Exchange and queue objects do not expose their name property, which makes it a bit awkward to refer to them in other areas, eg connection.publish(exchange.name, ..)
• Queue name is a required parameter, meaning auto generated queue names don't seem to be supported

Hope that's helpful - most of these are quite minor apart from the error handling one.

Hi,

I'm using amqp-coffee with reconnect = false option. Below is an example that reproduces the issue I faced recently. The example is a simple producer that publishes new message every second:

const AMQP = require('amqp-coffee');

const amqp = new AMQP({host: 'localhost', reconnect: false});

setInterval(() => {
  amqp.publish('coffee:test', '', 'hello', {}, (err) => {
    if (err) return console.log('Error: ' + err);
    console.log('Published');
  });
}, 1000);

The scenario is following:

1. run the example script
2. wait until at least one message is published
3. shut down message broker server

Expected result:
Option 1. For all next publishes the callback is invoked with an error
Option 2. The process crashes because of unhandled exception

Actual result:
Next publishes hang forever. Callbacks are never called. As I understand, publish waits until connection is back, but since reconnect is disabled it never happens.

What do you think?

I have same bug described in #23

Connection is closing after [heartbeat * 3] ms

Here is code which can reproduce bug on my localhost rabbit (mac) and server rabbit (ubuntu:
https://gist.github.com/vkfont/65d77cf461c075ebba60

Andreys-MacBook-Air:mservice-utils drago$ node debug/test
connected to rabbit@Andreys-Air v3.5.0
queue "amq.gen-xm_fKnPv_bzjKuguX8KsQQ" created
connection is closed: 30000ms
connected to rabbit@Andreys-Air v3.5.0
events.js:87
      throw Error('Uncaught, unspecified "error" event.');
            ^
Error: Uncaught, unspecified "error" event.
    at Error (native)
    at Consumer.emit (events.js:87:13)
    at Consumer._channelClosed (/Users/drago/Documents/ark/mservice-utils/node_modules/amqp-coffee/bin/src/lib/Consumer.js:183:14)
    at Consumer._channelClosed (/Users/drago/Documents/ark/mservice-utils/node_modules/amqp-coffee/bin/src/lib/Consumer.js:4:61)
    at Consumer.Channel._onChannelMethod (/Users/drago/Documents/ark/mservice-utils/node_modules/amqp-coffee/bin/src/lib/Channel.js:326:16)
    at Connection._onMethod (/Users/drago/Documents/ark/mservice-utils/node_modules/amqp-coffee/bin/src/lib/Connection.js:571:39)
    at AMQPParser.<anonymous> (/Users/drago/Documents/ark/mservice-utils/node_modules/amqp-coffee/bin/src/lib/Connection.js:4:61)
    at AMQPParser.emit (events.js:118:17)
    at AMQPParser.parseMethodFrame (/Users/drago/Documents/ark/mservice-utils/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:110:19)
    at AMQPParser.frameEnd (/Users/drago/Documents/ark/mservice-utils/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:80:16)

I have a problem with channels not closing. When using node-amqp I had the same issue, but resolved it with this queue option:

closeChannelOnUnsubscribe: true

Then when my socket.io client disconnected I would unsubscribe his consumerTag from the queue, and the channel would close:

socket.on('disconnect', function () {
    if (consumerTag) {
        my_queue.unsubscribe(consumerTag);
    }
});

Would there be an equivalent in amqp-coffee?

I'm trying all sorts of things like queue.channel.close(), consumer.cancel() and queue.delete(), but none seem to close the channels. I can see that they remain in rabbitmq dashboard with no consumer.

@barshow what was the design decision behind the 2 second channel inactivity timeout?

Hey,

I've been playing around with RPC modes & priority queues and noticed that there is a bug:

https://github.com/dropbox/amqp-coffee/blob/master/src/lib/Connection.coffee#L438

Specifically if value of the field is 0

  if args[field.name]

that check wont pass and there would be inconsistency between headers payload and actual payload (example, priority: 0). That being said we should simply check for hasOwnProperty to fix that issue

On the second note - https://www.rabbitmq.com/direct-reply-to.html - this is a nice feature to have, but require that publisher & consumer uses the same channel, which is currently not possible as they are always on the separate channels. I looked at how to do it myself, but didn't come to any conclusion - could be a decent idea to write an RPC channel with capabilities of both, but then I wouldnt want to duplicate a lot of code from Publisher/Consumer channels

What do you think is the best way to implement it?

Hello,

As we all know, node-amqp has plenty of shortcomings - mostly around the API quirks and general cruftiness due to age and back-compat - I'm interested in trying out alternatives like this.

However, there's no docs beyond the sample and the tests - are you guys working on some basic API docs at all?

if i will call queue.consumerCount or queue.messageCount durning connection problems - it stops fire callbacks, even if connection will back later, is something remain closed?

idea to fix:

• set flag to true on amqp ready event, and set to false on amqp close event
• return specific error in consumerCount/messageCount callback if flag is not active

CoffeeScript aside, it would be awesome if you could document in the README (or maybe blog post) why you decided to write this driver vs use the node-amqp driver. Including what issues this driver addresses that may have been over looked or in node-amqp.

Any idea what this means and how it can be fixed? Happens on reconnect attempt

12:31:58 arkmail-backend: connected to rabbit@ubuntu72 v3.5.3
12:31:58 arkmail-backend: connection is closed. Had error: { [Error: CHANNEL_ERROR - unexpected method in connection state running] replyCode: undefined }

And this would go on infinitely until I restart the process. Thanks for the help

Hi,

I'm using amqp-coffee with reconnect=false in one of my projects. During testing of some negative scenarios I faced a little bit odd behaviour. Here's a code example:

const amqp = new AMQP({host: 'localhost', reconnect: false});
amqp.consume('coffee:test:q', {}, (msg) => { /* ... */ });
amqp
  .on('error', (e) => console.log('ERROR: ' + e))
  .on('close', () => console.log('CLOSE'))

I run the program above and then manually take the local RabbitMQ server down. What I get in stdout is:

CLOSE
ERROR: Error: This socket has been ended by the other party
ERROR: Error: This socket has been ended by the other party
ERROR: Error: This socket has been ended by the other party
...

The Connection doesn't emit error only once, but continues emitting it every 10 seconds. As far as I see, it happens because of heartbeat timer which keeps triggering after the error. What I would rather expect is that once an error is emitted, the connection cancels all its internal timers.

What are your thoughts?

https://www.rabbitmq.com/connection-blocked.html

Not as useful as #52, but could be used to pause publisher or handle this scenario somehow

This isn't issue of this library, but I was wondering if you might have had similar issues and solved them already. I've already spent 6 hours researching and tuning anything that is tunable, but with no luck at all.

Issue is described here: rabbitmq/rabbitmq-server#564

TL;DR:
publisher - broker - consumer chain for 1 message takes 20ms on a VM and 0.5ms on my laptop

When we have the following options and #51 is merged in consumer gets new events

{
      host: 'localhost',
      port: 5672,
      login: 'guest',
      password: 'guest',
      vhost: '/',
      temporaryChannelTimeout: 6000,
      clientProperties: {
        capabilities: {
          consumer_cancel_notify: true,
        },
      },
},

Example log:

  amqp:AMQPParser 2 > method basicCancel {"consumerTag":"Vitalys-iMac.local-30767-1452811088079","noWait":1} +20s
  amqp:Consumer onMethod basicCancel, {"consumerTag":"Vitalys-iMac.local-30767-1452811088079","noWait":1} +1ms

It would be nice to have https://github.com/dropbox/amqp-coffee/blob/master/src/lib/Consumer.coffee#L196 extended with a handler for such an event (https://www.rabbitmq.com/consumer-cancel.html). It could simply emit an event, which a user must handle, or automatically try to reconnect. To synthetically generate such an event - create a consumer and then purge queue/exchange in the rabbitmq management console

Is this a sign something is wrong, or expected? The line in the stack trace that refers to my code points to empty space, which seems a little weird. The line right before it does message.ack(); in a consume callback.

(node) warning: possible EventEmitter memory leak detected. 11 listeners added. Use emitter.setMaxListeners() to increase limit.
Trace
    at Connection.EventEmitter.addListener (events.js:160:15)
    at Consumer.Channel (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/Channel.js:61:23)
    at new Consumer (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/Consumer.js:49:38)
    at ChannelManager.consumerChannel (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/ChannelManager.js:89:11)
    at Connection.consume (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/Connection.js:242:34)
    at /home/sgb/ui_client_adapter/lib/livestreamer.js:167:47
    at doneFn (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/Channel.js:240:11)
    at TemporaryChannel.Channel._onChannelMethod (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/Channel.js:315:37)
    at Connection._onMethod (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/Connection.js:538:39)
    at AMQPParser.<anonymous> (/home/sgb/ui_client_adapter/node_modules/amqp-coffee/bin/src/lib/Connection.js:4:61)

I got the following error after having a lot of rejects.
It may be that putting a message in a dead-letter queue adds things to the message header, and doing this repeatedly makes the message too large.

I need to be able to handle this kind of exception, but connection 'error' event is not firing on this kind of error.

2016-11-22T18:10:34+0000[ID=1] events.js:72
2016-11-22T18:10:34+0000[ID=1] throw er; // Unhandled 'error' event
2016-11-22T18:10:34+0000[ID=1] ^
2016-11-22T18:10:34+0000[ID=1] Error: Oversize frame 131318
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.error (/usr/local/logdog/card-guard-processor/node_
modules/amqp-coffee/bin/src/lib/AMQPParser.js:150:21)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.header (/usr/local/logdog/card-
guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:46:23)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.frameEnd (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:95:19)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.frame (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:64:21)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.header (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:51:21)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.frameEnd (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:95:19)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.frame (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:64:21)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.header (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:51:21)
2016-11-22T18:10:34+0000[ID=1] at AMQPParser.execute (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:31:33)
2016-11-22T18:10:34+0000[ID=1] at Socket. (/usr/local/logdog/card-guard-processor/node_modules/amqp-coffee/bin/src/lib/Connection.js:497:29)

Hi, I'd like to determine if a/all of the currently established connections to RabbitMQ are healthy to help assess the overall health of my application periodically. Can you give a recommendation to doing this?

One thought I had would be to use the existing client, kick off a heartbeat and listen for the response. However, I don't know how this would affect the amqp-coffee internals given that amqp-coffee would be performing its own healthchecks. Thoughts?

Awesome module by the way! 😍

lib-version: 0.1.22rc4
lang: JS

I am publishing a message to user queues through a topic exchange named "inbox.events".
The user queues are bound to the exchanges with routing keys following this scheme: "user.events.user_id.#".

Here is my publishing code:

This is the console output:

The problem is that although I am publishing to a different routing_key based on the current iteration of the users array, one user queue receives both messages. Are there any known issues with how published messages make their way out of the client lib and through to Rabbit?

Many thanks.

I'm getting an error when I try to call q.declare()

Error: Unmatched field {"name":"queue","domain":"shortstr"} at serializeFields (node_modules/amqp-coffee/bin/src/lib/serializationHelpers.js:261:19)

does this look familiar to anyone?

Has anyone tried to make this work with activeMQ instead of rabbit?
according to docs it should click, and my activeMQ instance is configured with:
<transportConnector name="amqp" uri="amqp://0.0.0.0:5672?maximumConnections=1000&amp;wireformat.maxFrameSize=104857600"/>

yet, if i try to run the basic.coffee from the examples (with port 5672 added), it just hangs around and i don´t see it appear as a consumer in my activemq or do anything useful until i kill it again.

amazon gift code or pizza vouchers on offer for any helpful pointers!

best regards
Andreas

Here is the output from NPM install after all of the module downloads. This is from installing v0.1.10. NPM install does work on a Mac.

[email protected] install /node_modules/amqp-coffee/node_modules/bson
(node-gyp rebuild 2> builderror.log) || (exit 0)

make: Entering directory /node_modules/amqp-coffee/node_modules/bson/build' CXX(target) Release/obj.target/bson/ext/bson.o SOLINK_MODULE(target) Release/obj.target/bson.node SOLINK_MODULE(target) Release/obj.target/bson.node: Finished COPY Release/bson.node make: Leaving directory/node_modules/amqp-coffee/node_modules/bson/build'

[email protected] install /node_modules/amqp-coffee
./scripts/compile.sh

amqp-coffee Compiling coffeescript to bin/
coffee> rm: missing operand
Try rm --help' for more information. npm ERR! [email protected] install:./scripts/compile.sh`
npm ERR! Exit status 123
npm ERR!
npm ERR! Failed at the [email protected] install script.
npm ERR! This is most likely a problem with the amqp-coffee package,
npm ERR! not with npm itself.
npm ERR! Tell the author that this fails on your system:
npm ERR! ./scripts/compile.sh
npm ERR! You can get their info via:
npm ERR! npm owner ls amqp-coffee
npm ERR! There is likely additional logging output above.

npm ERR! System Linux 2.6.32-279.19.1.el6.x86_64
npm ERR! command "/.nvm/v0.10.25/bin/node" "/.nvm/v0.10.25/bin/npm" "install" "amqp-coffee"
npm ERR! cwd
npm ERR! node -v v0.10.25
npm ERR! npm -v 1.3.24
npm ERR! code ELIFECYCLE
npm ERR!
npm ERR! Additional logging details can be found in:
npm ERR! /npm-debug.log
npm ERR! not ok code 0

Hi, I have a little program which publishes messages using amqp-coffee. I need to specify contentType of messages, because on a side of consumers there is a serializer which checks MIME type of a message and it should be 'plain\text'.

So, I tried to specify contentType: 'plain\text' on a conncetion.publish function using options argument but it always becomes 'string/utf8'. After reading code I found that in publish method of Publisher class it checks type of data and rewrites options.contentType.

It would be really cool, if this method skips checking type of data in cases when contentType already specified.

"An AMPQ 0.9.1..."

P.S. Not sure if this is the best way to get a typo fixed :)

I have setup Cluster on DinD PWD which creates and run cluster i am able to use management ui and see members in cluster but I am not able to connect to it there is no response for connection and client hangs.

version: "3"
services:
  rabbit1:
    image: lucifer8591/rabbitmq-server:3.7.17
    hostname: rabbit1
    ports:
      - "5672:5672"
      - "15672:15672"
    environment:
      - RABBITMQ_DEFAULT_USER=${RABBITMQ_DEFAULT_USER:-admin}
      - RABBITMQ_DEFAULT_PASS=${RABBITMQ_DEFAULT_PASS:-admin}
  rabbit2:
    image: lucifer8591/rabbitmq-server:3.7.17
    hostname: rabbit2
    links:
      - rabbit1
    environment:
      - CLUSTERED=true
      - CLUSTER_WITH=rabbit1
      - RAM_NODE=true
    ports:
      - "5673:5672"
      - "15673:15672"
  rabbit3:
    image: lucifer8591/rabbitmq-server:3.7.17
    hostname: rabbit3
    links:
      - rabbit1
      - rabbit2
    environment:
      - CLUSTERED=true
      - CLUSTER_WITH=rabbit1
    ports:
      - "5674:5672"

  const connectOptions = {
    host: [
      { host: 'ip172-18-0-40-bnf0r3vad2eg00bttg0g-5672.direct.labs.play-with-docker.com', port:5672 },
      { host: 'ip172-18-0-42-bnf0r3vad2eg00bttg0g-5673.direct.labs.play-with-docker.com', port: '5673' },
{ host: 'ip172-18-0-42-bnf0r3vad2eg00bttg0g-5674.direct.labs.play-with-docker.com', port: '5674' },

    ],
    login: "admin",
    password: "admin",
    heartbeat: 60,
    reconnect: true,
  }

  const connection = new AMQP(connectOptions);

connection.on('ready', () => console.log('Connected))
connection.on('error', () => console.log('error'))

Hi, can you do what the subject is describing? 👍

Hey. Great job on the wrapper. I'm facing an issue with exclusive consumers, however. When I set a consumer to exclusive, and try to run my workers from more than 1 machine, the second (and on) machine(s) throw errors on queue init. This makes sense, considering Rabbit throws an error when you try to connect to an already-consumed exclusive queue.

What's the right procedure here?

Hello,

After creating several consumers I'm getting the node's EventEmitter leak warning:

(node) warning: possible EventEmitter memory leak detected. 11 listeners added. Use emitter.setMaxListeners() to increase limit.
Trace
    at Connection.EventEmitter.addListener (events.js:160:15)
    at Consumer.Channel (/home/likewise-open/CORTEX/tsouza/workspaces/nodejs/stratos/node_modules/amqp-coffee/bin/src/lib/Channel.js:61:23)
    at new Consumer (/home/likewise-open/CORTEX/tsouza/workspaces/nodejs/stratos/node_modules/amqp-coffee/bin/src/lib/Consumer.js:47:38)
    at ChannelManager.consumerChannel (/home/likewise-open/CORTEX/tsouza/workspaces/nodejs/stratos/node_modules/amqp-coffee/bin/src/lib/ChannelManager.js:90:11)
    at Connection.consume (/home/likewise-open/CORTEX/tsouza/workspaces/nodejs/stratos/node_modules/amqp-coffee/bin/src/lib/Connection.js:259:34)

As a workaround, is it safe to call connection.setMaxListeners(Infinity) or the listeners will actually leak (i.e. on reconnections) ?

Regards

https://github.com/dropbox/amqp-coffee/blob/master/src/lib/Channel.coffee#L216-L247
https://github.com/dropbox/amqp-coffee/blob/master/src/lib/Queue.coffee#L32-L47

To replicate, first declare a HA+AD queue, then use another client and try creating HA queue with the same name. It should fail, but instead callback is returned successfully and then with an error

Can't seem to find anything in the code or the documentation about exchange to exchange routing, as described here: https://www.rabbitmq.com/e2e.html

Are there plans to implement this feature or would there be room for a contribution here?

"version": "0.1.22"

One of our servers crashed recently due to a type error, perhaps caused by a reimplementation of a topology upon reconnection.

Location: https://github.com/dropbox/amqp-coffee/blob/master/src/lib/Connection.coffee#L238

Stack "uncaughtException: TypeError: Cannot call method 'reset' of undefined\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/Connection.js:349:42\n at iterate (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:146:13)\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:157:25\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:251:17\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:154:25\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:248:21\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:612:34\n at Consumer.Channel._onChannelReconnect (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/Channel.js:190:14)\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/Channel.js:107:26\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:607:21\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:246:17\n at iterate (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:146:13)\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:157:25\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:248:21\n at /srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/async/lib/async.js:612:34\n at Consumer.Channel._onChannelMethod (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/Channel.js:330:37)\n at Connection._onMethod (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/Connection.js:602:39)\n at AMQPParser. (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/Connection.js:4:61)\n at AMQPParser.emit (events.js:106:17)\n at AMQPParser.parseMethodFrame (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:111:19)\n at AMQPParser.frameEnd (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:81:16)\n at AMQPParser.frame (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:64:21)\n at AMQPParser.header (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:51:21)\n at AMQPParser.execute (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:31:33)\n at Socket. (/srv/www/inboxmessenger/releases/20150722234551/node_modules/queue-api/node_modules/amqp-coffee/bin/src/lib/Connection.js:426:29)\n at Socket.emit (events.js:95:17)"

Thanks for maintaining the project. Could you release last version and tag it in git?

Would appreciate it a lot

This library can't connect if the server has limited channelMax.

We've recently gone live to production using your library to bridge RabbitMQ queues with clients who go on/off line quite frequently (think mobile). Whenever a client comes online we send a consume to the rabbitmq server and then when the client goes offline, we fire a consume.close(). During our beta stage things seemed to have gone just fine, however, recently, we've been seeing many open channels on our rabbitmq server that have no consumers. I was under the impression that a single channel would suffice for consuming for multiple queues where each queue ties to one client and one client only. Is my mental model incorrect?

Any thoughts?
Code looks somehting like this:

  var Conn = Rabbit.get();
   client.consumer = Conn.consume(/* params*/);
   ...
   client.on('disconnected', function(){
     client.consumer.close();
   });

iojs v2.4.0

mservice-arkmail-api-0 (err): RangeError: Attempt to allocate Buffer larger than maximum size: 0x3fffffff bytes
mservice-arkmail-api-0 (err):     at checked (buffer.js:174:11)
mservice-arkmail-api-0 (err):     at fromNumber (buffer.js:57:51)
mservice-arkmail-api-0 (err):     at new Buffer (buffer.js:42:5)
mservice-arkmail-api-0 (err):     at AMQPParser.header (/opt/arkmail-backend/node_modules/arkmail-backend/node_modules/mservice-utils/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:47:28)
mservice-arkmail-api-0 (err):     at AMQPParser.execute (/opt/arkmail-backend/node_modules/arkmail-backend/node_modules/mservice-utils/node_modules/amqp-coffee/bin/src/lib/AMQPParser.js:30:33)
mservice-arkmail-api-0 (err):     at Socket.<anonymous> (/opt/arkmail-backend/node_modules/arkmail-backend/node_modules/mservice-utils/node_modules/amqp-coffee/bin/src/lib/Connection.js:419:31)
mservice-arkmail-api-0 (err):     at emitOne (events.js:77:13)
mservice-arkmail-api-0 (err):     at Socket.emit (events.js:169:7)
mservice-arkmail-api-0 (err):     at readableAddChunk (_stream_readable.js:146:16)
mservice-arkmail-api-0 (err):     at Socket.Readable.push (_stream_readable.js:110:10)
mservice-arkmail-api-0 (err):     at TCP.onread (net.js:521:20)

I've been trying to integrate one, but with no success so far, main culprit is how reconnect is handled. basically what's done is called socket.connect() on a severed net socket, which reuses one, problem is that it doesnt work with tls (or I just dont know how to make it work). I wondered if it's better to abstract (re)connecting to reconnect-core module and setup proper handler there?

I've been working with these engines for a while now and it works pretty well, imo, npm constraints should include these version, since it's safe to use