Posted By:
Alison_Kakoschke
Posted On:
Tuesday, October 1, 2002 10:39 PM
Hi Andreas I have noticed the following behaviour in our system which is using swiftMQ. Two machines can be set up to send messages to each other successfully. We are then performing a test where we are simulating failure in the system by disconnecting and then reconnecting the network cable between these machines. In the meantime (during the period when the two machines are disconnected) one machine continues to publish messages to the other. What we notice is, that once the machines are reconnected, the second machine receives all messages sent to it during the disconnection state. We currently have set the time to live on all messages to 10 seconds when they are published, but they seem to arrive
More>>
Hi Andreas
I have noticed the following behaviour in our system which is using swiftMQ.
Two machines can be set up to send messages to each other successfully. We are then performing a test where we are simulating failure in the system by disconnecting and then reconnecting the network cable between these machines. In the meantime (during the period when the two machines are disconnected) one machine continues to publish messages to the other.
What we notice is, that once the machines are reconnected, the second machine receives all messages sent to it during the disconnection state. We currently have set the time to live on all messages to 10 seconds when they are published, but they seem to arrive at the other machine more than 10 seconds after they were published.
Is this correct behaviour? Is there any way that we can ensure messages die after 10 seconds regardless if the machines are currently in the connected or disconnected state? I have found an element called 'Listeners' on the JMS swiftlet which has an attribute - 'keepalive-interval'. Can this be used to control this feature?
Thanks in advance
Alison Kakoschke
Software Engineer
<<Less