I have a very simple SLSB written in EJB3. The packaged EAR file is copied to
"all/deploy" directory in two JBoss 4.0.4GA instances on two machines, n1 and
n2. These two JBoss instances form a cluster within one subnet with same out-of-box
configurations. The SLSB has annotation @Clustered. When looking up the SLSB, according to
the JBoss clustering manual, the application client uses
HA-JNDI:props.put(Context.PROVIDER_URL, "n1:1100,n2:1100")The app client worked
fine.
To simulate the failover, I undeployed myapp.ear from n1 by deleting it from
"all/deploy" on n1 machine. At this time, both JBoss instances are running, but
only n2 has my app deployed. Now executing the same app client, it hangs for about 30
seconds, then prints out "JNDI lookup fail" error.
If I use: props.put(Context.PROVIDER_URL, "n2:1100") the app client works.
If I stopped n1 instance, let the n2 instance running. The app client works with
props.put(Context.PROVIDER_URL, "n1:1100,n2:1100")
Before and after undeploying myapp.ear from n1, I looked at JMX console of the cluster.
Clicked on the jboss:service=DefaultPartition MBean. A list of IP addresses for the
cluster members (n1 and n2) is shown in the CurrentView field. So I think JGroup is
working fine here.
Are there any other configurations need to be changed?
Thanks,
James
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3965411#...
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...