The reason to have the connectionTimeout is to free up the thread to process
other requests, instead of blocking to see if the client is going to request
anything else. The default value is reasonable to serve up the page +
images etc. in one TCP connection. If you have a very high-traffic site,
then you usually want to disable keep-alive altogether (so the timeout value
(almost) doesn't matter, since Tomcat will serve one request only on the
thread). For a low-traffic site, it doesn't usually matter what you do :).
Also, this has nothing to do with the session timeout. It is possible that
if you leave your browser open when you go to lunch, that the session will
timeout but the request thread will still be sitting there waiting for the
browser to ask for something else.
Howard Jim [EMAIL PROTECTED] wrote in message
news:[EMAIL PROTECTED]
My current config looks like this:
Connector className=org.apache.coyote.tomcat4.CoyoteConnector
port=8011 minProcessors=5 maxProcessors=150
enableLookups=false redirectPort=443
acceptCount=10 debug=0 connectionTimeout=-1
useURIValidationHack=false
protocolHandlerClassName=org.apache.jk.server.JkCoyoteHandler/
I understand that the connection timeout is disabled with this setting. I
searched through the docs, etc, but all I ever found was set the connection
timeout in milliseconds.
My questions are, where would this matter? Why would I want to enable a
timeout here? If this is disabled, and connections are timed out elsewhere
(session, etc) shouldn't the processor still terminate normally?
I am assuming it is a safegaurd to prevent a run-away situation, but would
like some more input.
Thanks much,
Jim
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
