Build: GlassFish 9.1.1 B47
Windows 2003 Svr
I am running Richaccess longevity test on Windows 2003 Svr using Grizzly
Blocking Mode as the workaround in order for me to run my testing. This
workaround was given because of the Non-paged pool leak found in Windows 2003
I applied the same workaround with blocking=true on http-listener-1 and
increasing request-processing threads from 5 to 200. After blocking is enabled,
I am unable to access any of the instances in the cluster manually. On the
browser, the page continues to load for a while and after about 5 mins, it stops
loading but not does produce any error on browser nor the appserver logs. When I
turn off the blocking mode, the instances start to work.
This Blocking Mode was working for me till B40.
I have collected jstack information on instance3 during the time when the page
was trying to load. There are 2 jstack information taken at 1 min interval.
Attaching those files here.