Connection to STOMP broker is retried in a tight loop if an AMQP port is specified [SPR-11554]

[spring-projects-issues]

Sergi Almar opened SPR-11554 and commented

When specifying a wrong port as the relay port, connection is retried heavily and endlessly. This only happens when the port is open, for instance if you define the AMQP port instead of the STOMP port by mistake (or any other open port). This leads to the following logs:

13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-6] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-3] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-4] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-5] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-6] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-3] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-4] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-5] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-6] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}
13:57:23 [ringBuffer-ringbuffer-2] StompBrokerRelayMessageHandler - Established TCP connection to broker in session=stompRelaySystemSessionId
13:57:23 [reactor-tcp-io-3] StompEncoder - Encoded STOMP command=CONNECT headers={heart-beat=[10000,10000], host=[host], passcode=[pass], login=[user], accept-version=[1.1,1.2]}

---

Affects: 4.0.2

[spring-projects-issues]

Rossen Stoyanchev commented

The issue ended up getting resolved in the reactor project. I've tested with reactor-tcp 1.0.1.RELEASE, which is available immediately by referencing the http://repo.spring.io/libs-release repository. Later today it should also appear in Maven central.

[spring-projects-issues]

Rossen Stoyanchev commented

By the way the retry is not indefinite, it does stop eventually but it should be fixed now.