Well, each "run" of my test takes 25 minutes... so by the end of the second
run every session on the 1rst run should have finished a long time ago...
And even if that was the problem, how can we explain that the growth on the
replica is constantly increasing, but the growth of the "master" keeps
somewhat steady?
(If you check the last mail I send, it reports exactly that behaviour.  I am
assuming that in the master the memory usage is constant beacause it keeps
on expiring the http sessions and discarding them, but in the "replica"
there is no expiration and the sessions keep on memory forever...)

What do you think?

----- Original Message -----
From: "Bill Burke" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, September 17, 2003 4:20 PM
Subject: Re: [JBoss-user] Cluster "replica" absurd memory usage ?!?


> Maybe because you are creating new sessions every request and those
> sessions haven't timed out?  I think the default HTTP session timeout is
> 15 minutes.
>
> Joao Clemente wrote:
>
> > Ok, trying to figure out what is going wrong with my tests, I decided to
do
> > the following:
> >
> > I've runned my "100 user test" 10x. Test consists on about 15-20 user
> > actions (http requests) to the web/ejb application. Each action from a
> > specific user is separated by a minute, and users act in equal intervals
in
> > that minute. This means the 1rst action happens in the 1rst minute, 2nd
> > action in the second minute. This is a rate of 100/60 requests per
minute =
> > 1.6 requests per second.
> >
> > First I've runned the test with one machine, cluster settings enabled.
> > Memory usage before test started =  56Mb.
> > Memory usage after test end about 80Mb, peak mem usage was 89Mb.
> >
> > Then I've connected a second machine to the cluster. I restarted the
1rst
> > machine and cleaned up external application state.
> > Repeated the test directing requests to 1rst machine only. "Replica"
machine
> > did not got requests, only stood there as a cluster replica.
> > Memory usage before test started, around 56 Mb on each machine.
> > At 20% of the test (2 runs):
> >  1rst machine is using 99Mb, peak value is 105Mb.
> >  replica is using 225 Mb (!!!), and growing...
> > At 30% of the test (3 runs):
> >  1rst machine is using 103Mb, peak value is 105Mb.
> >  replica is using 285 Mb (!!!), and growing...
> >
> >
> > Now, who can explain:
> >  1 - why the 1rst machine memory usage increase? maybe related with
being
> > part of the cluster... but no session info needs to be replicated here
as
> > there is no requests being served by the replica machine...
> >  2 - why is the replica memory usage getting larger and larger each run
of
> > the test? It seems to grow about 60 Mb per run...


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.518 / Virus Database: 316 - Release Date: 11-09-2003



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user

Reply via email to