On 05/04/2012 22:01, Mark Eggers wrote: > Response at the bottom: > > /mde/ > >> ________________________________ >> From: Dhaval Jaiswal <dhaval.jais...@via.com> >> To: Tomcat Users List <users@tomcat.apache.org> >> Sent: Thursday, April 5, 2012 10:28 AM >> Subject: Re: SEVERE: Unable to start cluster. >> >> >> pfa server.xml >> >> The same i am using for all the 10 tomcats. >> >> >> >> >> On Thu, Apr 5, 2012 at 9:55 PM, Pid <p...@pidster.com> wrote: >> >> On 05/04/2012 15:13, Dhaval Jaiswal wrote: >>>> I have tested the below cluster manager; as per the instruction i have >>>> changed. However, when we try to deploy the apps on tomcat it was >>>> throwing error that >>>> >>>> WARNING: Unable to send map start message. >>>> Apr 5, 2012 1:37:12 PM org.apache.catalina.ha.session.BackupManager start >>>> SEVERE: Unable to start BackupManager >>>> java.lang.RuntimeException: Unable to start replicated map. >>>> at >>>> org.apache.catalina.tribes.tipis.AbstractReplicatedMap.init(AbstractReplicatedMap.java:230) >>>> >>>> Failed to start BackupManager: java.lang.IllegalArgumentException: >>>> Listener already >>>> exists:{}[org.apache.catalina.tribes.tipis.LazyReplicatedMap] >>>> >>>> >>>> >>>> i am using below server.xml for all the 10 tomcats. Do i have to change >>>> it for every tomcat. How would i fix it. >>>> >>>> >>>> >>>> <Cluster className="org.apache.catalina.ha.tcp.SimpleTcpCluster" >>>> channelSendOptions="6"> >>>> <Manager >>>> className="org.apache.catalina.ha.session.BackupManager" >>>> expireSessionsOnShutdown="false" >>>> notifyListenersOnReplication="true" mapSendOptions="6"/> >>>> >>>> <Channel >>>> className="org.apache.catalina.tribes.group.GroupChannel"> >>>> <Membership >>>> className="org.apache.catalina.tribes.membership.McastService" >>>> address="228.0.0.4" >>>> port="45563" >>>> frequency="500" >>>> dropTime="30000"/> >>>> <Receiver >>>> className="org.apache.catalina.tribes.transport.nio.NioReceiver" >>>> address="auto" >>>> port="4000" >>>> selectorTimeout="5000" >>>> timeout="60000" >>>> maxThreads="6"/> >>>> >>>> <Sender >>>> className="org.apache.catalina.tribes.transport.ReplicationTransmitter"> >>>> <Transport >>>> className="org.apache.catalina.tribes.transport.nio.PooledParallelSender" >>>> timeout="60000" >>>> keepAliveCount="0"/> >>>> </Sender> >>>> <Interceptor >>>> className="org.apache.catalina.tribes.group.interceptors.TcpFailureDetector"/> >>>> <Interceptor >>>> className="org.apache.catalina.tribes.group.interceptors.MessageDispatch15Interceptor"/> >>>> <!-- <Interceptor >>>> className="org.apache.catalina.tribes.group.interceptors.ThroughputInterceptor"/> >>>> --> >>>> </Channel> >>>> <Valve className="org.apache.catalina.ha.tcp.ReplicationValve" >>>> >>>> filter=".*\.gif|.*\.js|.*\.jpeg|.*\.jpg|.*\.png|.*\.htm|.*\.html|.*\.css|.*\.txt"/> >>>> <Valve >>>> className="org.apache.catalina.ha.session.JvmRouteBinderValve"/> >>>> <ClusterListener >>>> className="org.apache.catalina.ha.session.ClusterSessionListener"/> >>>> <ClusterListener >>>> className="org.apache.catalina.ha.session.JvmRouteSessionIDBinderListener"/> >>>> </Cluster> >>> >>> >>> Wow. That's pretty neat, I've never seen Tomcat start up with just the >>> cluster definition. >>> >>> How about you: >>> >>> 1) remove all comments >>> 2) remove all passwords >>> 3) post the whole server.xml inline in your next reply > > > Dhaval, > > > Please try to avoid posting in HTML. The font is sometimes difficult to read. > > Please don't cc me. I get the same message on the list, and with about 200 > mail messages per day on this account I might overlook stuff if I get a lot > of duplicate messages. > > Please post stuff inline and not as attachments. While I got the attached > server.xml file, the list generally strips attachments. > > All that being said, here are my comments: > > <!-- From your server.xml --> > <Listener > className="org.apache.catalina.core.AprLifecycleListener" SSLEngine="on" /> > <Listener > className="org.apache.catalina.core.JasperListener" /> > <Listener > className="org.apache.catalina.mbeans.ServerLifecycleListener" /> > <Listener > className="org.apache.catalina.mbeans.GlobalResourcesLifecycleListener" /> > > So you've removed the memory leak prevention listener from your > server.xml? Is there any reason why you've done that? > > <!-- from your server.xml --> > <Connector port="8080" protocol="HTTP/1.1" > connectionTimeout="20000" > redirectPort="8443" /> > > <Connector port="8009" > protocol="AJP/1.3" > redirectPort="8443" > backlog="100" > maxThreads="40000" /> > > No URIEncoding="UTF-8" on your connectors? I checked your web site and > there doesn't seem to be any characters that would be > impacted. However, I was thinking that setting UTF-8 on all connectors > is in general a good thing. > > 40000 threads for the AJP connector with only a backlog of 100? I hope that > > a) your Apache HTTPD server is configured appropriately > b) you can have this many file descriptors open > > <!-- from your server.xml --> > <Membership > className="org.apache.catalina.tribes.membership.McastService" > address="228.0.0.4" > port="45564" > frequency="500" > dropTime="30000"/> > > You said that you use this same server.xml for all 10 > Tomcats. However, in a previous message you said you were trying to > cluster 7 Tomcats (192.168.3.5 to 192.168.3.11). If you use the same > multicast address and port for all 10 Tomcats, then all 10 Tomcats > will be a part of the cluster. > >> I have tested the below cluster manager; as per the instruction i have > >> changed. However, when we try to deploy the apps on tomcat it was >> throwing error that > > Does this mean that the cluster actually starts without your web > application? If you've modified the logging.properties file, what does > the cluster logging file say? Please post the information inline as the > mailing list tends to strip attachments. > > The default dropTime is 3000 ms (3 seconds). Is your network busy > > enough so that a 30 second drop time is reasonable? If it is, then > maybe a better networking infrastructure is in order. > > <!-- from your server.xml --> > <Sender > className="org.apache.catalina.tribes.transport.ReplicationTransmitter"> > <Transport > className="org.apache.catalina.tribes.transport.nio.PooledParallelSender" > timeout="60000" > keepAliveCount="0"/> > </Sender> > > keepAliveCount="0" ? > > From the documentation: > > The number of requests that can go through the socket before the > socket is closed, and reopened for the next request. The default value > is -1, which is unlimited. > > So you're saying here that NO requests can go through the socket > before the socket is closed? At the best interpretation, you'll get > one message through the socket, then the socket closes and > reopens. This doesn't sound very performant. > > Other than that, your clustering setup does not look a lot different > than mine. I place my cluster in the <Host> element since I am > experimenting with farm deployment. Mine works even when changing from > the DeltaManager (only three node cluster) to the BackupManager. > > What is in your application's context.xml and web.xml files? Please > post both (inline - the list tends to strip attachments), with > comments and sensitive information removed. > > I suspect that you have a listener defined in either context.xml or > web.xml. > > All you really need to do to get session clustering to work is: > > 1. <distributable/> in web.xml > 2. Make sure that all session objects are serializalble. > > Actually there is a bit more, but the above are the basics. > > . . . . just my two cents. > /mde/
+1 to all that. p -- [key:62590808]
signature.asc
Description: OpenPGP digital signature