[opentaps-users] [opentaps - General Users Discussion] Opentaps Cluster [1.0]

[SourceForge.net <no...@so...>]

Read and respond to this message at: 
https://sourceforge.net/forum/message.php?msg_id=6281487
By: gjackson_000

I'm trying to enable clustering between two opentaps 1.0.2 instances. I'm using
apache with mod_jk infront of them. It is functioning fine with sticky sessions
enabled. I uncommented the cluster section in both opentaps instances and configured
them. They both launch without any exceptions and I see the following using
netstat:

tcp        0      0 192.168.0.30:4001       :::*                    LISTEN 
8099/java
udp        0      0 :::45564                :::*                           
8099/java

If I establish a session on node1 I see one Visit entry in the DB and session
is maintained. If I kill this node, then a new session is created on node2 with
a new Visit entry.

If I disable sticky sessions in mod_jk I get double entries in the Visit table
and it does not seem to maintain them. session is lost when switching protocols
(http, to https).

Any ideas to help? I'm open to any solution resulting in a active active
cluster.

Here is node1's cluster config:
 <property name="default-server-cluster" value="cluster">
                <property name="rep-valve-filter">
               
<property-value>.*\.gif;.*\.js;.*\.jpg;.*\.htm;.*\.html;.*\.txt;</property-value
>
                </property>
                <property name="manager-class"
value="org.apache.catalina.cluster.session.DeltaManager"/>
                <property name="debug" value="10"/>
                <property name="replication-mode" value="pooled"/>
                <property name="tcp-listen-host" value="192.168.0.30"/>
                <property name="tcp-listen-port" value="4001"/>
                <property name="tcp-sector-timeout" value="100"/>
                <property name="tcp-thread-count" value="6"/>
                <property name="mcast-addr" value="224.0.0.1"/>
                <property name="mcast-port" value="45564"/>
                <property name="mcast-freq" value="500"/>
                <property name="mcast-drop-time" value="3000"/>
            </property>
Node 2 is the same except a tcp-listen-host of 192.168.0.10

Here is the mod_jk workers def (the jvm route does match the node names):

workers.tomcat_home=/var/tomcat
workers.java_home=/usr/local/java
ps=/
worker.list=lb1,media1,jkstatus

worker.jvm1.port=9009
worker.jvm1.host=192.168.0.10
worker.jvm1.type=ajp13
worker.jvm1.lbfactor=1

worker.jvm2.port=9009
worker.jvm2.host=192.168.0.30
worker.jvm2.type=ajp13
worker.jvm2.lbfactor=1

worker.media1.port=8009
worker.media1.host=localhost
worker.media1.type=ajp13
worker.media1.lbfactor=1

worker.lb1.type=lb
worker.lb1.sticky_session=True
worker.lb1.balance_workers=jvm2,jvm1

What else do you need? Should I look for anything in the logs?

Thanks!!

______________________________________________________________________
You are receiving this email because you elected to monitor this forum.
To stop monitoring this forum, login to SourceForge.net and visit: 
https://sourceforge.net/forum/unmonitor.php?forum_id=487771