Petals CDK

Use case:
start an administration ESB with :
se-kpi started
se-notif started
then start a business ESB with :
se-jsr 181 with notification activated
then stop se-notif with the web console
then try to stop se-jsr 181 with the webconsole

this use case blocks on the se-jsr 181 stop half the time.

Possible causes :

• the se-notif unregister its endpoint on stop and then the admin ESB (which is the master) don't synchronize its registry with the business ESB (which is a slave) one. So the se-jsr 181 can send the exchange but never receive the response and blocks indefinitly on sendSync (cause sendSync without timeout).
• If I add a timeout, the problem is I don't know if the unregisterPublisher request will be processed by the se-notif when it restart. A possible solution could be to have a time to live parameter on the exchange. And fix its value to the same value as the timeout. Then you could manage timeout cases in both side of the exchange and not only on the consumer side like nowadays with the timeout parameter alone.