I have attempted to configure a tomcat cluster using version 5.5.7 of
tomcat.
I have configured a simple tcp cluster in the server.xml. I have two
tomcat instances installed on two different Linux servers
(Fedora core 3, IPTAPLES is turned off, there is not firewall).
I have apache and mod_jk set up on a third server. Multicast is
enabled!!!
When I start up each tomcat instance I receive the following message in
the catalina.out file.
------
Feb 23, 2005 9:21:49 AM org.apache.catalina.cluster.tcp.SimpleTcpCluster
start
INFO: Cluster is about to start
Feb 23, 2005 9:21:49 AM
org.apache.catalina.cluster.mcast.McastServiceImpl setupSocket
INFO: Setting multihome multicast interface to:/10.111.22.163
Feb 23, 2005 9:21:49 AM org.apache.catalina.cluster.mcast.McastService
start
INFO: Sleeping for 2000 secs to establish cluster membership
Feb 23, 2005 9:21:51 AM
org.apache.catalina.cluster.deploy.FarmWarDeployer start
INFO: Cluster FarmWarDeployer started.
Feb 23, 2005 9:21:51 AM org.apache.catalina.cluster.session.DeltaManager
start
INFO: Starting clustering manager...:/testapp
Feb 23, 2005 9:21:51 AM org.apache.catalina.cluster.session.DeltaManager
start
INFO: Manager[/testapp], skipping state transfer. No members active in
cluster group.
-----
The important thing to note here is that in the last line, I see
"INFO: Manager[/testapp], skipping state transfer. No members active in
cluster group."
At first I was thinking that multicast was not working properly.
However, I was researching the
Tomcat mailing list and found another thread discussing the same issue.
I tried to use the tool,
tomcat-replication.jar located on the CVS for tomcat. I can see the
multicast sessions being generated between
both servers using this tool. I also started up ethereal, the packet
sniffer. I noticed that both servers are generating multicast packets.
However after careful analysis of the packet information, I noticed that
my servers were looking for 127.0.0.1/4001 for the receiver port.
So I set the tcpListenAddress from "auto" to "10.111.22.163" This
seemed to fix the problem with the multicast packets sending out
the incorrect listener addresses. However, I am still getting the same
error...
"INFO: Manager[/testapp], skipping state transfer. No members active in
cluster group."
So here are my questions:
1. Is this really an error, or is session replication working and I am
not testing it properly?
2.What is the proper way to test that session replication is working
properly? Can anyone provide me with code that works. I am currently
using a
bean that I declare with <jsp:usebean scope=session.............Is this
wrong, should I use request.getAttribute(),
request.setAttribute()instead.
3.Can session replication replicate things like a class that I implement
as java.io.serializable?
Thanks in advance,
Randall
Cluster configuration from server.xml:
<Cluster
className="org.apache.catalina.cluster.tcp.SimpleTcpCluster"
managerClassName="org.apache.catalina.cluster.session.DeltaManager"
expireSessionsOnShutdown="false"
useDirtyFlag="true"
notifyListenersOnReplication="true">
<Membership
className="org.apache.catalina.cluster.mcast.McastService"
mcastAddr="224.0.0.5"
mcastPort="45564"
mcastFrequency="500"
mcastDropTime="3000"
mcastBindAddress="10.111.22.163"/>
<Receiver
className="org.apache.catalina.cluster.tcp.ReplicationListener"
tcpListenAddress="10.111.22.163"
tcpListenPort="4001"
tcpSelectorTimeout="100"
tcpThreadCount="6"/>
<Sender
className="org.apache.catalina.cluster.tcp.ReplicationTransmitter"
replicationMode="pooled"
ackTimeout="15000"/>
<Valve
className="org.apache.catalina.cluster.tcp.ReplicationValve"
filter=".*\.gif;.*\.js;.*\.jpg;.*\.png;.*\.htm;.*\.html;.*\.css;.*\.txt;
"/>
<Deployer
className="org.apache.catalina.cluster.deploy.FarmWarDeployer"
tempDir="/tmp/war-temp/"
deployDir="/tmp/war-deploy/"
watchDir="/tmp/war-listen/"
watchEnabled="false"/>
</Cluster>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
