Very strange.. I looked at the patch, I think we should spend more time on that, the sleep doesn't look good and I'm afraid we may have some problems with the CS.
The worker was supposed to be thread safe, like a servlet - and so far it works in most cases. But all thread problems are difficult to trace, and if we can't find the real problem I'm ok with using the pool - but with a different patch if possible. I run tests with jk2 and it seems to work fine ( -c 100). Costin On Fri, 21 Jun 2002, Mladen Turk wrote: > > > > -----Original Message----- > > From: GOMEZ Henri [mailto:[EMAIL PROTECTED]] > > Sent: 21. lipanj 2002 18:36 > > > > Status/Stability : > > ---------------- > > > > I stressed latest jk 1.2.0 with ab using 127 > > concurrents requests, asking for 1 Million requests > > for the good old HelloWorldExample servlet > > (Mladen discovered a problem with ab with dynamic > > content-length) > > > > I didn't have any error now, with any > > combinaison of Apache 1.3/2.0 and > > Tomcat 3.3.1/4.0.4. > > > > What about those forced closed connections, and not closed sockets? > > > > > We need now people to test jk 1.2.0 against > > apache 1.3/2.0 on windows (and windows/cygwin ?). > > Also we need testing for iplanet. > > > > Apache 2.0.39+ with Tomcat 4 is ok on WIN2K, but has problems on XP. It > would be great if someone can run that on the other XP machine, but on > the other hand doesn't really matter because of afd.sys bug on XP (MSKB > article Q317949). > > > MT. > > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > > -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
