users@glassfish.java.net

Re: Very Odd behavior of LifeRay on Glassfish 3.0.1

From: Pete Helgren <pete_at_valadd.com>
Date: Fri, 28 Sep 2012 12:00:53 -0600

OS is Red Hat Enterprise Linux Server release 5.6 (Tikanga) x86_64 It
has 10GB RAM and two processors. JVM is configured for 4GB.
Performance seems OK, no bottlenecks I can see.

It looks like there is an http-thread-pool that has 1000 threads and a
second thread pool thread-pool-1 that has 1000 threads as well. Not
sure if there is anything else I need to configure. I am going to check
out some performance and configuration best practices for Glassfish as
well as Apache.

Just guessing at the db. I guess Apache could be an issue as well. I
am looking at all possible causes at this point.

Thanks for the info.

Pete Helgren
Value Added Software, Inc
www.petesworkshop.com
GIAC Secure Software Programmer-Java

On 9/28/2012 11:11 AM, Roel_D wrote:
> What is your OS?
>
> And with 50k visitors, how many glassfish theads do you use?
>
> I run some selfwritten portlets that at some point create 500+ simultanious mysql connections but haven't seen your behaivior.
>
> If you can test a copy of your webserver then i would suggest to try neoload. They have extremely good software to stresstest your webserver. ( i have once tested it and pretty impressed. I have no connection with them)
> You can then maybe find the causes without interupting user experience.
>
> Kind regards,
>
> The out-side