Skip to content

Incorrect Recover commands during shutdown with spring-rabbit >= 3.0.11 #2941

@marcogramy

Description

@marcogramy

I'm using Spring Boot 3.2.12 (which includes spring-rabbit 3.1.8 and spring-amqp 3.1.8) and I'm using several SimpleMessageListenerContainer consumers.

During consumer shutdown, specifically during connection closure, I noticed unusual Recover commands that I hadn't observed before. By analyzing the network traffic (using tcpdump) on port 5673, I captured the following message exchange, which is unexpected:
Image

After conducting several tests with previous versions of the spring-rabbit and spring-amqp libraries, I found that downgrading to spring-rabbit-3.0.10 resolves the issue. The network traffic in this scenario shows the expected behavior:
Image

Starting from spring-rabbit-3.0.11 onwards, I observe the problematic Recover commands again.
My RabbitMQ server version is 4.0.5.

I suspect this behavior is related to changes introduced in spring-rabbit-3.0.11 concerning shutdown procedures, as documented in issue #2594 and the corresponding commit: 02db80b.

Can you help me understand why this behavior occurs?
Thank you in advance for your assistance.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions