Currently, org.jnp.interfaces.NamingContext caches NamingServerStubs and retries the Naming.lookup() call MAX_RETRIES number of times. MAX_RETRIES is set to 10 as part of its declaration and is non-configurable as far as I know.
Is there a way to control the number of retries made by org.jnp.interfaces.NamingContext.lookup() when the Naming.lookup() call fails?
Following is a brief description of the scenario when this would be useful (to me anyway):
- A plain old java object tries to lookup a remote ejb using org.jnp.NamingContextFactory as its JNDI service provider.
- It succeeds the first time. Which means that a NamingServerStub for the remote machine has been cached by org.jnp.interfaces.NamingContext.
- Unplug the remote machines network cable.
- Try to lookup the same ejb again. An exception is thrown from the NamingContext after 10 retires for lookup(). This takes 3-4 minutes.
- I need to reduce that length of time.
- Is there a way that can be done by configuring some properties?
- Other options would be to manage the connection timout myself via threads or direct tcp/ip calls. Is there another option?
Thanks.
Varun Rupela.