[rabbitmq-discuss] Distributed garbage collection

[Steven Taylor]

Hi Guys,

I'm digging into RabbitMQ, and would like to build in fault
tollerance, scalability, and garbage collection within the cloud (highly
asynchronous context).  Probably the solution should be message queue based,
and heuristic based.  eg. I lost the connection, so self destruct.  The
queue is filling up, so fire up more instances. Talk to AWS and conduct an
audit on nodes sending in a heart beat. Destroy anything that does not
respond.

Would anyone care to share some experiences?  There have got to be a few
papers out there on this.  I'm hoping to get through this without rebuilding
the operating system.  Given that RabbitMQ relies on Erlang, I'm hoping to
get some of this for free, and that new nodes can make lightweight
considerations in code.

thanks,
-Steven
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20101008/91393021/attachment.htm>