Hi, We are using Glassfish 3.1.1 CA in our production environment. It has
following configuration. Total servers = 3 (Solaris x86) RAM each server = 12
GB Virtual CPU each server = 6 Glassfish Instance on each server = 2 (2
instance in one cluster same machine, no cross server cluster). Each
Server.... MAX HTTP thread pool = 500 MAX Threadpool -1 = 200 MAX EJB
Container POOL = 150 Application details: Application is EJB 3.1 using
Asynchronous threads for parallel processing. One main thread invoked through
Web service (HTTP) and it register the request (DB call) and then spawn 2
parallel asynchronous threads for parallel processing. This is mandatory for
us as we have to respond to web service request quickly for reducing response
time and parallel processing is bit complex and takes bit time. In production
with peak load (around 40-70 concurrent request per second), the main request
gets served, however the 2 parallel asynchronous threads takes 20 mins, to 1
hour to start processing. Kindly suggest. Thanks in advance. Ashish
--
[Message sent by forum member 'ashishkp']
View Post: http://forums.java.net/node/891263