The error has nothing to do with the messages or fields in them. It is
saying that a client connection is trying to use the same AMQP
container-id as another existing client connection. The broker treats
the connection container-id as the ClientID in JMS terms, only
allowing one connection to use a given value at a time.
On Thu, 23 Feb 2023 at 22:56, George Sexton
<geor...@mhsoftware.com.invalid> wrote:
>
> I’m using ActiveMQ, Broker Engine Version 5.15.6 on AWS. The instance
> size is mq.t2.micro, in Single-instance broker deployment mode.
>
> A Golang application that has been working for years started
> misbehaving. The calling application is getting error messages:
>
> time="2023-02-23 21:35:17.1885" level=error msg="Retrying send"
> app=addef error="*Error{Condition: amqp:invalid-field, Description:
> Broker: xxx-xxx-xxxx - Client: 6bcc2aa9-468c-4c38-a16e-8713e6e7833b
> already connected from tcp://99.99.99.99:34834, Info:
> map[invalid-field:container-id]}" id=6bcc2aa9-468c-4c38-a16e-8713e6e7833b
>
> and:
>
> time="2023-02-23 21:35:17.1887" level=*error*msg="failed to handle ct
> deletion event" app=addef ctid=04f7519f-1125-4047-7047-c475d0ee87b2
> *error*="failed to send ct changed message: **Error*{Condition:
> amqp:invalid-field, Description: Broker: xxx-xxx-xxxx- Client:
> 6bcc2aa9-468c-4c38-a16e-8713e6e7833b already connected from tcp://
> 99.99.99.99::34834, Info: map[invalid-field:container-id]}"
> request_id=42558916-d800-485e-9abb-a2ffd156a366
>
> When I initially looked at the system, the home page showed Store
> Percent Used to be 100%. I purged the dead-letter queue, and now Store
> Percent Used shows 11%. After the purge, I rebooted the instance, but
> the application is still failed.
>
> I turned on Cloudwatch logging for AmazonMQ, and I can see the startup
> messages. From looking at the AWS article on enabling logging, INFO is
> the default log level and can’t be changed. Aside from startup, all I
> see is a few entries:
>
> Transport Connection to: tcp://10.100.16.40:33962 failed:
> org.apache.activemq.transport.InactivityIOException: Channel was
> inactive for too long |
> org.apache.activemq.broker.TransportConnection.Transport |
> AmqpInactivityMonitor Async Task:
> java.util.concurrent.ThreadPoolExecutor$Worker@47ec0444[State = -1,
> empty queue
>
> I looked at the application code, and a message from the dead-letter
> queue, and can say there is not a container-id field in them. I searched
> around and found some references to invalid-field container-id, but they
> seem to be internal references:
>
> https://issues.apache.org/jira/browse/AMQ-5591
>
> Does anyone have any ideas on how I can troubleshoot this, or areas I
> can look into? Thanks for helping!
>
>
> --
> George Sexton
> (303) 438 9585 x102
> MH Software, Inc.
