Problem:
- Product Name: VisiBroker
- Product Version: 6.0
- Product Component: Naming Service
- Platform/OS Version: ALL
When a client side ORB runtime tries to fail-over to the alternate object in the cluster by
calling select() on the cluster, it gets this exception:
Read: com.inprise.vbroker.GIOP.ReplyMessage[version=1.2, message_size=51,
request_id=44, moreFragments=false, service_contexts={}, reply_status=USER_EXCEPTION]
exception com.inprise.vbroker.CosNamingExt.Empty {}
at com.inprise.vbroker.CosNamingExt.EmptyHelper.read(EmptyHelper.java:32)
at com.inprise.vbroker.CosNamingExt._ClusterStub.select(_ClusterStub.java:55)
at com.inprise.vbroker.naming.NamingBindInterceptor.bind_failed(NamingBindInterceptor.java:140)
at com.inprise.vbroker.interceptor.ChainBindInterceptorImpl.bind_failed(ChainBindInterceptorImpl.java:76)
...
What does this exception mean?
Resolution:
The definition of this exception in the Visibroker API doc is stated as this:
" This exception is raised by the select() and destroy() method when the cluster
is empty. "
This points to the fact that the cluster currently does not have any objects
registered to it.
One of the possible reasons could be that if Naming Service cluster-failover or
master-slave configuration is used, the naming graph on either Name Services are
out of sync, and could be due to the difference in the instances of the backing
stores between the 2 Name Services.
#VisiBroker
#Security
