http://issues.apache.org/jira/browse/AXIS2-935

I was wondering if anyone had come across a recommended way to work around
the issues listed in this bug. Right now, this causes our JVMs to hang. 

In all cases, it seems recommended from a performance and memory standpoint
to cache the ConfigurationContext object, but this issue is more or less a
showstopper if you plan to use either the single or multithreaded HTTPSender
objects for calling your Axis2 server instance.

We basically get a hung JVM once we've crossed some threshold either tied to
pooled httpclient objects or open file descriptors, that sometimes clears
itself. (after an unidentified period of time)

Ideally, it would be great if we could definitively answer this question:
If REUSE_HTTP_CLIENT, is there a way to get rid of the CLOSE_WAIT file
descriptors other than garbage collect?

-- 
View this message in context: 
http://www.nabble.com/Axis2-Client%2C-JVM-Hangs--Recommended-stopgap-solution-for-JIRA-AXIS2-935-tf3430423.html#a9563019
Sent from the Apache SOAP - User mailing list archive at Nabble.com.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to