Hello Herbert -
On Thu, 12 Oct 2000, Herbert Kornfeld wrote:
> I'd like to test radiator (2.16.3 on hp-ux 11, perl5.6. authbyldap2)
> under heavy load scenarios, to help plan for servers. What I'd like to
> do is have some data that allows me to plot average response time vs the
> number of simultaneous requests. (Say with the latter ranging through
> 2,5,10,50,100,1000,10000,50000 or until radiator breaks ;))
>
> I've looked at radpwtst, but whilst -iterations and -time are useful,
> they serialise the requests, whereas I want to simulate nnnn NASes all
> sending requests at the same time.
>
> What is the best way to do this?
>
> I could write a driver script that started nnnn copies of radpwtst in
> the background, each one sending xxxxx iterations and timing the
> results, but this leads to other issues (like, nnnn processes competing
> amongst each other for resource. Maybe the sticky bit would help, like
> in the old days?)
>
Just set up Radiator on one or two additional machines and run multiple copies
of radpwtst on each one against the Radiator host. Thats what we do here for
our own stress testing.
hth
Hugh
--
Radiator: the most portable, flexible and configurable RADIUS server
anywhere. SQL, proxy, DBM, files, LDAP, NIS+, password, NT, Emerald,
Platypus, Freeside, Interbiller, TACACS+, PAM, external, etc, etc.
Available on Unix, Linux, FreeBSD, Windows 95/98/2000, NT, MacOS X.
===
Archive at http://www.starport.net/~radiator/
Announcements on [EMAIL PROTECTED]
To unsubscribe, email '[EMAIL PROTECTED]' with
'unsubscribe radiator' in the body of the message.
