[AMQ-5231] Failover Transport timeout option causes connection failures in some cases where it shouldn't

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 5.8.0, 5.9.0, 5.9.1, 5.10.0
• Fix Version/s: 5.10.1, 5.11.0
• Component/s: Transport
• Assignee: Timothy A. Bish
• Reporter: Timothy A. Bish
• Created: 18/Jun/14 16:08
• Updated: 15/Dec/14 22:01
• Resolved: 18/Jun/14 17:01

Description

Originally added in AMQ-2061 the timeout option is used to cause a send of Message that's gone out async to fail if the connection doesn't come back in some amount of time.

The problem is that the option is currently applied to everything that goes through FailoverTransport oneway() and this can cause a Connection start where the broker is not up to fail regardless of the maxReconnectAttempts or startupMaxReconnectAttempts value configured for the transport.

We need to refine the timeout logic to only apply to Message instances and not to other commands like ConnectionInfo so that a Connection start honors the normal failover transport reconnect configuration logic.

Issue Links

• is related to: AMQ-2061 Timeout on failover transport without interruption of retries
• relates to: AMQCPP-545 Failover transport timeout options should only operate on Message sends.