tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From János <janos.l...@yale.edu>
Subject Failing over is failing when tomcats are on different machines
Date Mon, 04 Aug 2008 21:13:56 GMT
Hi,

I setup a three node cluster on my Mac and tested a fail over with  
three nodes successfully.  My next step was to setup tomcat on my  
Windows 2000 machine too and migrate one of the nodes over there.    
After completing the setup I tried fail over again, but it is not  
working.  The two nodes are failing over to each others on my Mac,  
but tnever to the third node on the windows machine.  The Apache2 web  
server is running on the Mac and I use mod_jk to connect.

In the three server.xml file I just added an additional Host.  Here  
is the Engine portion from node2 on the Mac

...
     <Engine name="Catalina" defaultHost="bml0042.yalepath.org"  
jvmRoute="node2">
                 <Cluster  
className="org.apache.catalina.ha.tcp.SimpleTcpCluster"/>
                 <Realm  
className="org.apache.catalina.realm.UserDatabaseRealm"
              resourceName="UserDatabase"/>
                 <Context path="" docBase="ROOT"  
className="org.apache.catalina.ha.context.ReplicatedContext"/>
                 <Host name="bml0042.yalepath.org"  appBase="webapps"
             unpackWARs="true" autoDeploy="true"
             xmlValidation="false" xmlNamespaceAware="false"/>
                 <Host name="bml0039.yalepath.org"  appBase="webapps"
             unpackWARs="true" autoDeploy="true"
             xmlValidation="false" xmlNamespaceAware="false"/>
     </Engine>
....

here is my workers.properties file from the conf directory of apache2:

...
worker.list = lbnodes,jkstatus

worker.lbnodes.type=lb
worker.lbnodes.balance_workers=node1,node2,node3

worker.jkstatus.type=status

worker.node1.type = ajp13
worker.node1.host = 10.84.2.42
worker.node1.port = 8009
worker.node1.lbfactor = 1
worker.node1.redirect=node2
worker.node2.type = ajp13
worker.node2.host = 10.84.2.42
worker.node2.port = 8109
worker.node2.lbfactor = 1
worker.node2.redirect=node3
worker.node3.type = ajp13
worker.node3.host = 10.84.2.39
worker.node3.port = 8209
worker.node3.lbfactor = 1
worker.node3.redirect=node1

worker.node1.sticky_session = True
worker.node2.sticky_session = True
worker.node3.sticky_session = True

worker.node1.sticky_session_force = False
worker.node2.sticky_session_force = False
worker.node3.sticky_session_force = False
....

The node3 on the Windows machine participates in the Cluster and  
registers the other nodes, and even the session replication is  
working, it is just the fail over that has the problem.
I do not want to run an IIS or even Apache2 on Windows, I want only  
Tomcat there.  It looks to me that although the workers.properties  
are there but somehow JK knows only about the first two nodes on the  
Mac and not the third one on the Windows 2000 machine.

Do have to put FQDN-ed values into jvmRoute like  
bml0039.yalepath.org:/node3, or similar as Engine parameters ??  Do I  
have to share out C;\node3 as a share ??

Any good advise, or pointers to the appropriate files where a cluster  
is working on different machines with different OS-es?

Thanks ahead,

János



---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Mime
View raw message