Hi there,
We are running a JBossMQ server that is both accessible over https using jbossmq-httpil, and accessible locally over jnp. Our client wants high availability so we are trying to figure out how to deploy a failover scheme.
We have been reading http://wiki.jboss.org/wiki/Wiki.jsp?page=JBossMQHAOverview, but it is not quite clear how this works for HTTPS?
We now have an Apache server with mod_jk in front of our JBossMQ and we need to use the Apache server's external name for the service:
<mbean code="org.jboss.mq.il.http.HTTPServerILService"...
<attribute name="URLHostName">node.external.name</attribute>
...
</mbean>
Mod_jk is now configured as:
JkMount /invoker jbossmq
JkMount /invoker/* jbossmq
JkMount /jbossmq-httpil jbossmq
JkMount /jbossmq-httpil/* jbossmq
worker.jbossmq.port=8009
worker.jbossmq.host=node.internal.name
worker.jbossmq.type=ajp13
How should we configure multiple
HTTPServerILService mbeans in a cluster to make them different, since they all need to be accessed through the same Apache server?
I guess we also need to deploy a HA-JNDI service on all nodes, but does the http-invoker service work with a HA-JNDI?
And how should we configure the client to acces the HA-JNDI over https:
java.naming.provider.url=https://node.external.name:443/invoker/JNDIFactorySSL
java.naming.factory.initial=org.jboss.naming.HttpNamingContextFactory
context.lookup("HTTPConnectionFactory");
Should we configure the client to use multiple http url's?
Or could we use Apache load balancing to access the JNDI's and do the failover?
Do we need HA-JNDI then?
Many thanks in advance,
Geert.
