laylaylilom
asked on
Tomcat 4.1.24 CoyoteConnector Stops responding
During high volume usage Tomcat's CoyoteConnector stops responding. WHY?
I have a web site that gets about 500,000 page views/day. The current architecture is on Win2K Server Tomcat 4.1.24 + mod_jk 1.2.4 + Apache 1.3.28. During the peak times of the usage the Coyote Connector stops responding, and I cannot find the reason. Driving me mad.
I know that tomcat is not crashing because the process is still alive and it also responds to the calls made thru port 8080 the http connector.
I also know that it is not the database that causes the error, because I have other app servers that can use the database.
I think the problem is not caused by running out of threads since the same problem occurs even when the maxProcessors is set to an absurd number like 9000.
I am sure that the bottleneck is not apache, because even though I open up the connector to accept connections from all IP addresses, when the connector stops responding and apache starts giving 500 errors, I cannot connect to the port 8009. Therefore, I am pretty sure the source of the problem is in the Connectors. WHY?!?!? PLEASE HELP!
The current connector configuration:
<Connector className="org.apache.coyo
address="127.0.0.1" port="8009" minProcessors="50" maxProcessors="1256"
enableLookups="true" redirectPort="8443"
acceptCount="300" debug="0" connectionTimeout="20000"
useURIValidationHack="true
protocolHandlerClassName="
PS: I also tried to have 4 of these connectors running from different ports with modjk loadbalancing, however, they all go down one by one.
I have a web site that gets about 500,000 page views/day. The current architecture is on Win2K Server Tomcat 4.1.24 + mod_jk 1.2.4 + Apache 1.3.28. During the peak times of the usage the Coyote Connector stops responding, and I cannot find the reason. Driving me mad.
I know that tomcat is not crashing because the process is still alive and it also responds to the calls made thru port 8080 the http connector.
I also know that it is not the database that causes the error, because I have other app servers that can use the database.
I think the problem is not caused by running out of threads since the same problem occurs even when the maxProcessors is set to an absurd number like 9000.
I am sure that the bottleneck is not apache, because even though I open up the connector to accept connections from all IP addresses, when the connector stops responding and apache starts giving 500 errors, I cannot connect to the port 8009. Therefore, I am pretty sure the source of the problem is in the Connectors. WHY?!?!? PLEASE HELP!
The current connector configuration:
<Connector className="org.apache.coyo
address="127.0.0.1" port="8009" minProcessors="50" maxProcessors="1256"
enableLookups="true" redirectPort="8443"
acceptCount="300" debug="0" connectionTimeout="20000"
useURIValidationHack="true
protocolHandlerClassName="
PS: I also tried to have 4 of these connectors running from different ports with modjk loadbalancing, however, they all go down one by one.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.