[rabbitmq-discuss] what would cause queue size to immediately go to 0?

[Dave Syer]

On Nov 5, 4:07 pm, Shane Witbeck <sh... at digitalsanctum.com> wrote:
> Stepping through the code (ChannelN) I saw that noack is true with code
> unchanged. I set container.setChannelTransacted(true); and now noack is
> false and I see the desired behavior.

Cool.  That's what we expect.

> I should mention that I'm using the AmqpOutboundEndpoint in the Spring
> Integration sandbox to publish messages from a JdbcInboundAdapter. Should I
> be aware of any other changes that would effect this setup?

Nothing in the current snapshot, but I am working on
https://jira.springframework.org/browse/AMQP-69 (it's more than a test
now) to support best efforts 1PC (http://www.javaworld.com/javaworld/
jw-01-2009/jw-01-spring-transactions.html).  By analogy with the
Spring JMS support this will be enabled if you inject a transaction
manager into the message listener container.  This might be useful to
you if you want to use a DataSourceTransactionManager in your Spring
Integration flow, and make it single threaded - then if there are any
failures (except a rabbit broker failure after the DB commit) your
system returns to its original state before the transaction.