On 04/09/2010 10:19 AM, David Zhou wrote:
Hi all, i have set up a cluster of C++ broker A and B, and brokers in the cluster are synced by corosync. i can send a message to broker A and retrieve it from broker B. my question is: Is there any automated way of failover for python client? (for java or c++ client, there seems to have a FailoverManager to handle this problem and we just supply the FailoverManager an url that specifies some candidate brokers.) But for python client, in the example illustrated in the doc, i haven't see a similar case of failover handling with FailoverManager. If there is none, does anyone have a hint for me to implement failover with python client?
There is a newer api that supports automatic reconnect and replay; that is in the messaging module. There have been a few changes around some method names for that api recently but it is settling down now. What version are you using, or are you working from trunk?
(Btw there is an equivalent api in c++ now that aims to make the experience using these two clients more uniform. The failover support in that is nicer in my view than the 0-10 specific api.)
--------------------------------------------------------------------- Apache Qpid - AMQP Messaging Implementation Project: http://qpid.apache.org Use/Interact: mailto:[email protected]
