The goal of this project is to illustrate an issue at startup when trying to setup RabbitMQ processing.
$ cargo run --bin producer
$ cargo run --bin consumer
Sometimes, the consumer does not consume the messages that are already present in the queue.
- run the producer a few times (for example 3 times):
cargo run --bin producer - run the consumer (
cargo run --bin consumer). It is supposed to print Got message on primary queue. with the details of the message.
In many situation, the consumer get stuck at Stream on primary-queue and does not process the message (even though the RabbitMQ admin panel tells us they are there).
An interesting thing is that the RabbitMQ admin tells us that one message is
unacked in the primary-queue. So the consumer has fetched it but it never
goes farther and the Got message... is never printed.
You need to run the consumer a few times to get it to work. It looks like some threading (starving? deadlock?) issue.
If the consumer manage to start and to process the first message, there seems that it will process new messages as expected and won't be stuck again.
The producer simply publish a message in the primary-queue.
The consumer creates two channels, one to interact with the primary-queue and
one to interact with the secondary-queue. It basic_consume() the
primary-queue, simulate some work on the message, send the result to the
secondary-queue and then basic_ack() the initial message from the
primary_queue.