The connectionTimeout was the solution, increasing the value of the other attributes simply delayed the problem.
I had been using the Ajp13Connector earlier, then noticed this new attribute on they CoyoteConnector so tried that as a work around. So far so good in my testing.
-Simon
Milt Epstein wrote:
On Wed, 18 Dec 2002, Simon Chatfield wrote:I made the following change <Connector className="org.apache.coyote.tomcat4.CoyoteConnector" port="8009" minProcessors="5" maxProcessors="100" enableLookups="true" redirectPort="8443" acceptCount="100" debug="0" connectionTimeout="20000" useURIValidationHack="false" protocolHandlerClassName="org.apache.jk.server.JkCoyoteHandler"/>Could you indicate what you changed? maxProcessors? acceptCount? connectionTimeout? (Perhaps you posted what it was earlier in the thread, but if so I missed it.) Thanks. FWIW, my comment below was about Ajp13Connector, I didn't realize you were using CoyoteConnector. And looks like Ajp13Connector has the maxProcessors and acceptCount attributes, but not connectionTimeout.And it appears to have fixed the lock-up problem at least. The problem wasn't in tomcat in my estimation, the apache/mod_jk side was holding onto the connection and not allowing the next request to use it. The timeout kills off the connections that are being held onto by apache and so no lockups. I think this is only a temporary kludge/fix, but for the other people who were having the same problem, try this for now. Thanks for everyone's help. -Simon ps. I also upgraded from 4.1.12 to 4.1.17 in my debugging process though that didn't solve the problem in and of itself. Milt Epstein wrote:On Wed, 18 Dec 2002, Simon Chatfield wrote:The number of open connections.... # netstat -a |grep 8009 |wc -l 175 # actually, I think that shows both sides of the connection between apache and tomcat + the listener, so that's 87 connections...[ ... ] I think how many of these can exist is controlled by the maxProcessors attribute setting in the Ajp13Connector Connector tag in server.xml. Also, you can see messsages about their starting up in cataling_log.YYYY-MM-DD.txt (where YYYY-MM-DD is the date). And it may be correct that they are not closed/stopped once they're started. Milt Epstein Research Programmer Integration and Software Engineering (ISE) Campus Information Technologies and Educational Services (CITES) University of Illinois at Urbana-Champaign (UIUC) [EMAIL PROTECTED] -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>-- Simon Chatfield The Chatfield Group email: [EMAIL PROTECTED] phone: 602-971-9598 web: http://www.thechatfieldgroup.com -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>Milt Epstein Research Programmer Integration and Software Engineering (ISE) Campus Information Technologies and Educational Services (CITES) University of Illinois at Urbana-Champaign (UIUC) [EMAIL PROTECTED] -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
-- Simon Chatfield The Chatfield Group email: [EMAIL PROTECTED] -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
