tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ivan Vasquez" <i...@itos.uga.edu>
Subject RE: Clustering and session replication
Date Thu, 30 Sep 2004 15:49:51 GMT
Filip,

We will file a bug. In the meantime we will try 5.0.25 as it appears it
doesn't have the problem. 

Thanks a lot.
Ivan.

-----Original Message-----
From: Filip Hanik - Dev [mailto:devlists@hanik.com] 
Sent: Thursday, September 30, 2004 11:39 AM
To: Tomcat Users List
Subject: Re: Clustering and session replication

ok, so when the context is reloaded you are saying it doesn't work.
sounds like that upon a context reload the new context does not
get a ClusterManager, and instead gets the StandardManager assigned.
File a bug in bugzilla, and we will take care of it

We never reload our contexts in production so that is why I haven't seen
what you are seeing

Filip
----- Original Message -----
From: "Ivan Vasquez" <ivan@itos.uga.edu>
To: "Tomcat Users List" <tomcat-user@jakarta.apache.org>
Cc: "Brian Lawrence" <lawrence@itos.uga.edu>
Sent: Thursday, September 30, 2004 10:05 AM
Subject: RE: Clustering and session replication


Filip,

To discard any problems in application settings, we copied an entire
application from production (clustered tomcat 5.0.25) to the cluster
that's having problems (5.0.28). We verified context settings and load
balancing.

The first time the application is used after a tomcat restart, it works
well (load balancing and session replication). But after reloading the
application in any node (i.e. context reloads), the log will show the
warnings and timeout errors that I described before. Session replication
fails from that point on.

>>From the little I understand, I think clustering is doing its job, the
problem is it cannot find the "context manager" for that application
(despite both nodes have <distributable/> in web.xml). However, the
context is properly configured because we can access dbcp pools and mail
connections defined within.

Ivan.

-----Original Message-----
From: Filip Hanik - Dev [mailto:devlists@hanik.com]
Sent: Wednesday, September 29, 2004 5:14 PM
To: Tomcat Users List
Subject: Re: Clustering and session replication

let me reiterate

org.apache.catalina.cluster.tcp.SimpleTcpCluster messageDataReceived
WARNING: Context manager doesn't exist:/nahrgisdev

that means there is no clustered context under this name.
If it works from server 1 to 2 but not from 2 to 1, then you need to
look over your configuration again.

The good news is it worked for you in one direction, you have to focus
on the above error message,
that message and that message alone is the reason the requesting server
times out

Filip

----- Original Message -----
From: "Ivan Vasquez" <ivan@itos.uga.edu>
To: "Tomcat Users List" <tomcat-user@jakarta.apache.org>
Sent: Wednesday, September 29, 2004 3:54 PM
Subject: RE: Clustering and session replication


Filip,

Thanks for your reply. Each app has identical web.xml files:

<?xml version="1.0" encoding="ISO-8859-1"?>

<!-- Required in order to have default EL support -->
<web-app xmlns="http://java.sun.com/xml/ns/j2ee"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://java.sun.com/xml/ns/j2ee web-app_2_4.xsd"
    version="2.4">

  <distributable/>

  <!-- General description of your web application -->
  <display-name>NAHRGIS Clustered</display-name>
  <description>
    NAHRGIS GIS Application
  </description>

...

The context, with datasources and other resources, works as expected
(but it lives in conf/Catalina/localhost/nahrgisdev.xml, not in
server.xml).

The cluster is set as follows:

<Cluster className="org.apache.catalina.cluster.tcp.SimpleTcpCluster"
  managerClassName="org.apache.catalina.cluster.session.DeltaManager"
  expireSessionsOnShutdown="false"
  useDirtyFlag="false">

<Membership
  className="org.apache.catalina.cluster.mcast.McastService"
  mcastAddr="228.0.0.4"
  mcastPort="45564"
  mcastFrequency="500"
  mcastDropTime="3000"/>

<Receiver
  className="org.apache.catalina.cluster.tcp.ReplicationListener"
  tcpListenAddress="10.0.0.1"
  tcpListenPort="4001"
  tcpSelectorTimeout="100"
  tcpThreadCount="6"/>

<Sender
className="org.apache.catalina.cluster.tcp.ReplicationTransmitter"
                replicationMode="pooled"/>

<Valve className="org.apache.catalina.cluster.tcp.ReplicationValve"
filter=".*\.gif;.*\.js;.*\.jpg;.*\.htm;.*\.html;.*\.txt;"/>

<Deployer className="org.apache.catalina.cluster.deploy.FarmWarDeployer"
                      tempDir="/tmp/war-temp/"
                      deployDir="/tmp/war-deploy/"
                      watchDir="/tmp/war-listen/"
                      watchEnabled="false"/>
</Cluster>

After further testing, one node consistently times out upon reload:

requesting session state from
org.apache.catalina.cluster.mcast.McastMember[tcp://10.0.0.2:4001,10.0.0
.2,4001
...
SEVERE: Manager[/nahrgisdev], No session state received, timing out

While the other one gets it in 100 or so ms.

Respectfully,
Ivan.




-----Original Message-----
From: Filip Hanik - Dev [mailto:devlists@hanik.com]
Sent: Wednesday, September 29, 2004 4:17 PM
To: Tomcat Users List
Subject: Re: Clustering and session replication

org.apache.catalina.cluster.tcp.SimpleTcpCluster messageDataReceived
WARNING: Context manager doesn't exist:/nahrgisdev


this is telling you that the server that spits out this message does not
have a <distributable/> context configured by the name
"nahrgisdev"

all the servers in the cluster must have the all the "<distributable/>"
webapps configured identically

Filip

----- Original Message -----
From: "Ivan Vasquez" <ivan@itos.uga.edu>
To: <tomcat-user@jakarta.apache.org>
Sent: Wednesday, September 29, 2004 3:03 PM
Subject: Clustering and session replication


We are having some problems with a cluster of 2x Tomcat 5.0.28 (JDK
1.4.2) regarding session replication. The problem manifests after a
distributable application is reloaded. From that point on, a successful
login to our application causes the same page to be reloaded by the
other node (mod_jk is load balancing) as if no progress was being done.
A new session ID is generated every time by either node.

Every time it happens, the log shows:

org.apache.catalina.cluster.tcp.SimpleTcpCluster messageDataReceived
WARNING: Context manager doesn't exist:/nahrgisdev

On application reload, a node shows this error

Sep 29, 2004 3:47:26 PM org.apache.catalina.cluster.session.DeltaManager
start
WARNING: Manager[/nahrgisdev], requesting session state from
org.apache.catalina.cluster.mcast.McastMember[tcp://128.192.104.11:4001,
128.192.104.11,4001, alive=253140]. This operation will timeout if no
session state has been received within 60 seconds
Sep 29, 2004 3:48:26 PM org.apache.catalina.cluster.session.DeltaManager
start
SEVERE: Manager[/nahrgisdev], No session state received, timing out.

However, we can telnet the other side on that port. Internal NICs are
not configured for the cluster because initially we thought they were
the problem.

The application's context is stored in
conf/Catalina/localhost/nahrgisdev.xml.

Of course, the application works well in a non-clustered environment.

Thanks for any advice,
Ivan.

---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org

---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org


Mime
View raw message