[rabbitmq-discuss] node-amqp: { [Error: COMMAND_INVALID - second 'channel.open' seen] code: 503 }

[Chris]

Thank you, Michael.  The Tracer is a very cool tool.  Thanks for pointing
it out!  It definitely confirms the issue regarding multiple opens:

1363291338032: <Tracer-0> ch#8 -> {#method<channel.open>(out-of-band=),
null, ""}
1363291338032: <Tracer-0> ch#8 -> {#method<channel.open>(out-of-band=),
null, ""}
1363291338070: <Tracer-0> ch#8 -> {#method<channel.open>(out-of-band=),
null, ""}
1363291338070: <Tracer-0> ch#8 -> {#method<channel.open>(out-of-band=),
null, ""}

Now I still just have to figure out how/why they are being sent!

-Chris


On Thu, Mar 14, 2013 at 10:08 AM, Michael Klishin <
michael.s.klishin at gmail.com> wrote:

>
> 2013/3/14 Chris <stuff at moesel.net>
>
>> First, the non-node-specific question:  Does anyone have any pointers on
>> how to debug such an issue?  I've already looked at logs and found which
>> channel is the problem-- but if I comment out the code that creates that
>> channel, then it just fails on a different channel the next time (btw-- you
>> don't explicitly create channels in node-amqp, they are created for you
>> when you subscribe to queues, declare exchanges, etc.).
>>
>
> Using Wireshark or RabbitMQ tracer [1].
>
>
>>
>> And the more node-specific question:  RabbitMQ docs seem to indicate it
>> is usually an error in the client (node-amqp in our case), but I wonder if
>> perhaps we are making some newbie node-amqp mistake that might lead to this
>> 503 error?  Is there anyone with node-amqp experience that has been through
>> this before?
>>
>
> This issue indicates a race condition when a library assumes it need to
> automatically open a channel after channel.open
> has been sent. Or it may be because channel ids are not incremented
> properly in the client so two channels end up
> having the same id.
>
> Finally, race conditions can originate in your own code but it's
> impossible to tell without
> seeing the code.
>
> 1. See at the bottom of http://www.rabbitmq.com/api-guide.html
> --
> MK
>
> http://github.com/michaelklishin
> http://twitter.com/michaelklishin
>
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20130314/32ad3d95/attachment.htm>