You should try my approach: A think time of 0.000 seconds. In 4 separate processes.
Let's just say I will try not to do that again. :-) Also, maybe we need some rate limiting. Is there a way that is usually done in Lift that allows for efficient processing (so that thousands of unauthorized requests don't bring down the server). Apparently even checking for session validity is too much in these situations. Ethan On Fri, Dec 18, 2009 at 12:11 AM, Markus Kohler <[email protected]> wrote: > I wouldn't run through. I have to work on error handling. I also used a > think time of 0.5 to 1.0 seconds. > So I can probably generate more load if necessary. > Markus > > On Dec 18, 2009 4:47 AM, "Richard Hirsch" <[email protected]> wrote: > > Just checked the Stax logs - Ethan's microblogging-DOS client hit us harder > :-> > > Memory just went up to 133 and we had a peak of 250 requests yesterday > - a little worrying is that request time peaked at 148s. We have to > check this out in more detail. With Ethan's killer hacker tool, we had > a peak of over 4200 requests! > > We still haven't deployed ESME in "production mode" yet on this server > - so I'll do that today. > > D. > > On Thu, Dec 17, 2009 at 9:26 PM, Markus Kohler <[email protected]> > wrote: > ... just to warn... >
