I experienced a port shutdown on Tomcat 3.2B6 when I held my finger down on the refresh key in IE. Apparently the number of aborted connections generated by the browser was greater than Tomcat could handle. Port 8080 shutdown, while 8007 stayed up. I enabled SYN flooding protection in my Linux firewall and the problem went away. Not sure if this is your problem or not, but its worth a look. Chris -----Original Message----- From: jblum [mailto:[EMAIL PROTECTED]] Sent: Thursday, November 09, 2000 12:46 PM To: [EMAIL PROTECTED] Subject: RE: Help! Tomcat 3.2 BETA 6 crashes with only 20 simulated users I have some additional useful info about our problem below. We ran the same test as below except we eliminated Apache and mod_jk totally from the equation by using port 8080 in the URL to connect directly with Tomcat. And after a few hours Tomcat stopped responding with no errors in the logs. The Tomcat processes still stayed around. Any help on this would be very appreciated. Thanks Jim Blum > -----Original Message----- > From: jblum [mailto:[EMAIL PROTECTED]] > Sent: Tuesday, November 07, 2000 3:54 PM > To: [EMAIL PROTECTED] > Cc: Jim Blum > Subject: Help! Tomcat 3.2 BETA 6 crashes with only 20 simulated users > > > I am using Tomcat version 3.2 Beta 6 on a Solaris Ultra Enterprise 2 SUN4U > box running SUNOS 5.7 with 1 gigabyte of ram and 2 processors > with plenty of > disk space and 70 megabytes of swap space. > > When we run Seque's silkperformer version 4.02 webtesting on a Windows PC > with 20 simultaneous users (simulated browsers) hitting the same > URL on the > Tomcat server SUN box, (the test waits for a response and then repeats the > URL) Tomcat no longer responds. Nothing else is running on the SUN box > except for apache version 1.3 which is passing the URL to tomcat > via mod_jk. > I am using JSP and 2 custom tags which I wrote myself. > > There are 4 or 5 tomcat processes running after Tomcat stops responding to > the URL with an Apache HTTP/1.1" 500 620 response error. > > The only change I have made is I have added the following line to the > tomcat.sh file: > TOMCAT_OPTS="-Xmx96m" to set the memory size for JVM to use. > > I am running java version "1.2.1" > Solaris VM (build Solaris_JDK_1.2.1_04, native threads, sunwjit). > > There are no apparent errors in the Apache log files except for the > HTTP/1.1" 500 620 errors. > > In looking at the mod_jk log file, I find the following error: > > [jk_ajp12_worker.c (121)]: Into jk_endpoint_t::service > [jk_connect.c (108)]: Into jk_open_socket > [jk_connect.c (115)]: jk_open_socket, try to connect socket = 8 > [jk_connect.c (124)]: jk_open_socket, after connect ret = -1 > [jk_connect.c (143)]: jk_open_socket, connect() failed errno = 146 > [jk_ajp12_worker.c (134)]: In jk_endpoint_t::service, sd = -1 > [jk_ajp12_worker.c (152)]: In jk_endpoint_t::service, Error sd = -1 > [jk_ajp12_worker.c (163)]: Into jk_endpoint_t::done > > I can't tell whether this error is happening because tomcat has > already quit > responding or whether there is some of out of sockets resource problem. > > Any help on this problem would be appreciated, as we obviously can't use > Tomcat until this problem is resolved. > > Thanks > Jim Blum > Talk City Inc. > {λΏ~} >
