Sizing of WebLogic Cluster Servers

I have an application that needs to be implemented in BEA WebLogic 5.1 (
JSP, Servlet, EJB) and deployed on Solaris 8 talking to Oracle
8.0/Solaris (in clustered environment).
I am in the process of sizing the application. Following are some of the
basic sizing information:
No of users: 6000
Usage Pattern: High during evenings in US time zone
Concurrent users: 300
EJBs types: 200 entity and 50 session beans ( stateless), 4 stateful
session beans
State Per User: 2K
Database Tables: 100
No of screens: 80
Data Accessed per User Request: 20K
Cluster: Yes
No of Machines in Cluster: 2
Clustering Type: Recommended 2- tier ( with Web Server on a different
machine)
What kind of SUN machine should I be buying?
I mean, ES 220 or ES 450 ...or a higher end machines? What should be the
swap space, RAM, no. of processors?
Any existing documentation or specification will be very valuable.
Thanks,
Hari

Hi,
First of all weblogic server has ONLY one Admin Sever per Domian which should be used for administrative work, hence we cannot have have a backup of Admin server.
Weblogic Server Architecture
Yes, the best way is to use horizontal clustering as shown below where all the managed servers are in a cluster.
Machine-1
- Admin Server
- MS-1
- MS-3
Machine-2
- MS-2
- MS-4
HTTP Servers Architecture
Assuming you would be using Apache HTTP server, hence using weblogic Apache plugin you would be able to communicate with weblogic servers. In hhtpd.conf file you have to load the module and in if-module block you would have to give all the information about the weblogic server which would then try to connect with weblogic servers as well as will load-balance between them.
For more information check the below link for Apache HTTP server.
Topic:Installing and Configuring the Apache HTTP Server Plug-In
http://download.oracle.com/docs/cd/E15051_01/wls/docs103/plugins/apache.html#wp131399
Topic:General Parameters for Web Server Plug-Ins
http://download.oracle.com/docs/cd/E12840_01/wls/docs103/plugins/plugin_params.html#wp1143055
Hope this helps you.
Regards,
Ravish Mody

Similar Messages

How to setup weblogic cluster in 6.1 SP4

          Hi All,
          I have a Licensed weblogic in my solaris servers. I need to test weblogic clustering
          . How can i simulate weblogic clustering is there is any easy steps to do that
          and what should i do in order to make my EJB`s cluster aware. And some ppl say
          that i need to take a seperate weblogic cluster license other than the weblogic
          license.Kindly help me in proceeding further
          Thanks in advance
          Bala.J


Hi
I have same problem at beforetime.
You can do follow:
Step 1:
copy rt.jar form JDK 1.2.1\jre\lib\jar
to J2EE1.2.1\lib\system\
Step 2:
Set J2ee_classpath = c:\j2ee1.2.1\lib\system\rt.jar
Step 3:
Restart computer and J2ee
You can try it
Help can help you
David

Having issue with start weblogic cluster with tangosol cluster

Hi,
Oracle Coherence Version 3.3.1/389p1
Grid Edition: Development mode
We are using Weblogic 8.1.5 with Tangosol 3.3.1 on Linux servers.
And we added the initializing logic in the servlet's init() method to get all NamedCaches and put into the ServletContext.
When we start weblogic cluster, the first weblogic member will startup successfully with following messages :
<Nov 7, 2007 10:12:30 AM EST> <Info> <HTTP> <BEA-101047> <[2007-11-07 10ServletContext(id=259640596,name=clusterqa,context-path=)] initObjects: init>:12:31.565 Oracle Coherence 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): Loaded operational configuration from resource "zip:/home/server/clusterqa/wls81/DOCVIEW/docqa1/.wlnotdelete/extract/docqa1_DOC_clusterqa/jarfiles/WEB-INF/lib/coherence.jar!/tangosol-coherence.xml"
2007-11-07 10:12:31.598 Oracle Coherence 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): Loaded operational overrides from file "/home/www/WEB-INF/lib/tangosol-coherence-override.xml"
Oracle Coherence Version 3.3.1/389p1
Grid Edition: Development mode
Copyright (c) 2000-2007 Oracle. All rights reserved.
2007-11-07 10:12:31.938 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): Loaded cache configuration from file "/home/www/WEB-INF/lib/pub-search-cache-config.xml"
2007-11-07 10:12:31.983 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): sun.misc.AtomicLong is not supported on this JVM; using a synchronized counter. Though safe to ignore, you may upgrade to BEA's 1.5 JVM to fix this issue.
2007-11-07 10:12:33.267 Oracle Coherence GE 3.3.1/389p1 <Warning> (thread=Main Thread, member=n/a): UnicastUdpSocket failed to set receive buffer size to 1428 packets (2096304 bytes); actual size is 89 packets (131071 bytes). Consult your OS documentation regarding increasing the maximum socket buffer size. Proceeding with the actual value may cause sub-optimal performance.
2007-11-07 10:12:34.118 Oracle Coherence GE 3.3.1/389p1 <D5> (thread=Cluster, member=n/a): Service Cluster joined the cluster with senior service member n/a
2007-11-07 10:12:37.508 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Cluster, member=n/a): Created a new cluster with Member(Id=1, Timestamp=2007-11-07 10:12:33.323, Address=10.5.176.86:8088, MachineId=48982, Edition=Grid Edition, Mode=Development, CpuCount=4, SocketCount=2) UID=0x0A05B056000001161AAB782BBF561F98
2007-11-07 10:12:37.736 Oracle Coherence GE 3.3.1/389p1 <D5> (thread=Invocation:Management, member=1): Service Management joined the cluster with senior service member 1
2007-11-07 10:12:38.168 Oracle Coherence GE 3.3.1/389p1 <D5> (thread=DistributedCache, member=1): Service DistributedCache joined the cluster with senior service member 1
<Nov 7, 2007 10:12:38 AM EST> <Info> <HTTP> <BEA-101047> <[ServletContext(id=259640596,name=clusterqa,context-path=)] xslProcessor: init>
But trying to start the second weblogic member server, the startup process is stucked after tangosol cache initialization and the second weblogic member server never up running. Please see following messages :
<Nov 7, 2007 9:49:38 AM EST> <Info> <HTTP> <BEA-101047> <[ServletContext(id=153019550,name=clusterqa,context-path=)] initDSNames: init>
<Nov 7, 2007 9:49:42 AM EST> <Info> <HTTP> <BEA-101047> <[ServletContext(id=153019550,name=clusterqa,context-path=)] initObjects: init>
2007-11-07 09:49:43.156 Oracle Coherence 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): Loaded operational configuration from resource "zip:/home/server/clusterqa/wls81/DOCVIEW/docqa2/.wlnotdelete/extract/docqa2_DOC_clusterqa/jarfiles/WEB-INF/lib/coherence.jar!/tangosol-coherence.xml"
2007-11-07 09:49:43.188 Oracle Coherence 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): Loaded operational overrides from file "/home/www/WEB-INF/lib/tangosol-coherence-override.xml"
Oracle Coherence Version 3.3.1/389p1
Grid Edition: Development mode
Copyright (c) 2000-2007 Oracle. All rights reserved.
2007-11-07 09:49:43.528 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): Loaded cache configuration from file "/home/www/WEB-INF/lib/pub-search-cache-config.xml"
2007-11-07 09:49:43.571 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Main Thread, member=n/a): sun.misc.AtomicLong is not supported on this JVM; using a synchronized counter. Though safe to ignore, you may upgrade to BEA's 1.5 JVM to fix this issue.
2007-11-07 09:49:44.829 Oracle Coherence GE 3.3.1/389p1 <Warning> (thread=Main Thread, member=n/a): UnicastUdpSocket failed to set receive buffer size to 1428 packets (2096304 bytes); actual size is 89 packets (131071 bytes). Consult your OS documentation regarding increasing the maximum socket buffer size. Proceeding with the actual value may cause sub-optimal performance.
2007-11-07 09:49:45.419 Oracle Coherence GE 3.3.1/389p1 <D5> (thread=Cluster, member=n/a): Service Cluster joined the cluster with senior service member n/a
2007-11-07 09:49:45.555 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Cluster, member=n/a): Failed to satisfy the variance: allowed=16, actual=47
2007-11-07 09:49:45.555 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Cluster, member=n/a): Increasing allowable variance to 19
2007-11-07 09:49:46.040 Oracle Coherence GE 3.3.1/389p1 <Info> (thread=Cluster, member=n/a): This Member(Id=2, Timestamp=2007-11-07 09:49:45.69, Address=10.5.176.85:8088, MachineId=48981, Edition=Grid Edition, Mode=Development, CpuCount=4, SocketCount=2) joined cluster with senior Member(Id=1, Timestamp=2007-11-07 09:45:10.205, Address=10.5.176.86:8088, MachineId=48982, Edition=Grid Edition, Mode=Development, CpuCount=4, SocketCount=2)
Could you please explain why it happens, and what should I do to resolve this issues ?
Many Thanks,
Bing

Hi, Gene
Thank you for the response. I will send you our full log files and thread dumps.
I just want to give you more details about our cases :
1. This only happened without starting cache servers (com.tangosol.net.DefaultCacheServer).
2. And our application which are running on weblogic cluster will just call "CacheFactory.getCache("XXX")", and running as the Tangosol DataClient.
3. All weblogic member servers will be up running successfully if our cache servers are up running.
Also I tried to test another case :
Suppose all weblogic instances and cache server instances are up running. Now I trying to restart (kill weblogic instance process and restart) one of the weblogic member, It will up running successfully only if add some sleep times after killing weblogic processes and restarting it. Looks like tangosol cluster need certain time to aware the member has left cluster, then the restart process will be successful.
Questions :
1. Should we start our weblogic cluster only after cache server cluster is up running ?
2. How do we decide how many time we should wait before start new process to join the cache cluster ?
Could you please help to explain this one for me and let us if there anyway we can do to avoid the problem.
Many Thanks !!!
Bing

Single sign-on in weblogic cluster?

          I am using wls 7.0 sp1 that form a cluster by two managed servers each at different
          physical machine.
          In front of the servers, a hardware load-balancer providing a virtual ip that
          mapped to the real ip of the 2 machines.
          I deployed a web application to the cluster. The html & jsp files are all protected
          by Form Based Authorization.
          My expectation is, whenever I sign-on to the first managed server, the login credential
          should also be valid for the second managed server.
          But it is not the case.
          My situation is:
          - open brower with url: http://virtual.ip:7001/admin.jsp
          - the load balancer map the request to my first managed server.
          - the response from the first managed server prompt the login page for username
          and password.
          - after POST the information, my target admin.jsp page display properly.
          - when I click the link within the admin.jsp page, eg. http://virtual.address:7001/status.jsp
          there is two possibilities:
          1. the load balancer route the request to the first managed server and everything
          well.
          or
          2. the load balancer route the request to the second managed server and response
          with login page again.
          The second case is definitely not what I expected.
          Question:
          Is it the default behavior of Weblogic cluster on the single sign-on? I think
          my login is against a WLS domain (at least a cluster), not to a specific web server!
          Any comments are welcome.


          Note my answer is not directly related to your question, but it might help.
          From my experience to configure WL cluster with hardware load balancer, the session
          stick (layer 7 switch) should be used (either using the session cookie, or ServerID
          cookie). That is, unless one instance is down, all requests from the same session
          are better go to the same server. I ever tried round-robin or random load balance
          policy at hardware load balancer without session stick, and the result was as
          disappointing as you have seen.
          "franky c." <[email protected]> wrote:
          >
          >
          >I am using wls 7.0 sp1 that form a cluster by two managed servers each
          >at different
          >physical machine.
          >In front of the servers, a hardware load-balancer providing a virtual
          >ip that
          >mapped to the real ip of the 2 machines.
          >I deployed a web application to the cluster. The html & jsp files are
          >all protected
          >by Form Based Authorization.
          >
          >My expectation is, whenever I sign-on to the first managed server, the
          >login credential
          >should also be valid for the second managed server.
          >
          >But it is not the case.
          >
          >My situation is:
          >- open brower with url: http://virtual.ip:7001/admin.jsp
          >- the load balancer map the request to my first managed server.
          >- the response from the first managed server prompt the login page for
          >username
          >and password.
          >- after POST the information, my target admin.jsp page display properly.
          >- when I click the link within the admin.jsp page, eg. http://virtual.address:7001/status.jsp
          > there is two possibilities:
          > 1. the load balancer route the request to the first managed server
          >and everything
          >well.
          > or
          > 2. the load balancer route the request to the second managed server
          >and response
          >with login page again.
          >
          >The second case is definitely not what I expected.
          >
          >Question:
          >Is it the default behavior of Weblogic cluster on the single sign-on?
          >I think
          >my login is against a WLS domain (at least a cluster), not to a specific
          >web server!
          >
          >
          >Any comments are welcome.
          >
          >

Patchset3 (11.1.1.4/wls 10.3.4) Remote debugging JDPA in weblogic cluster

Hi,
I am using Patchset3 (JDeveloper 11.1.1.4/ weblogic server 10.3.4). I was wondering how i can remote debug with
a) weblogic managed server that is part of a weblogic cluster
b) the cluster itself that contains multiple managed servers.
Note that i do understand how to remote debug against a weblogic server (e.g. http://jdeveloperfaq.blogspot.com/2009/12/faq-4-how-to-remote-debug-adf.html), but have not been able to find out how i can achieve this against a cluster that has multiple managed servers. Any directions will be appreciated.
Thanks,

Hi,
haven't tried it but here are some suggestions that might be of help:
1) Each managed server for each cluster node should be setup for remote debugging. This means that each managed server must be started with the -Xdebug -Xnoagent -Xrunjdwp:transport=dt_socket,address=4000,server=y,suspend=n* options in the WebLogic start-up script. Also Tunneling must be enabled for each managed server for each cluster node.
2) Make sure that you specify a different connection port for each managed server. This is indicated by the address* specifier above.
3) It seems logical to me that you will need to start separate instances of JDeveloper - one for each managed server that you need to attach to.
Let us know how it turns out,
Nick

Round Robin doesn't work in Bea Weblogic Cluster

hello!
          we are using a weblogic (V5.1 SP8) cluster in the following configuration:
          2 Weblogic App-Servers as Servlet Containers.
          4 Weblogic App-Servers as EJB Containers.
          if a client connects to the first servlet - server, round robin between the
          4 EJB-servers behind doesn't work. It seems that the servlet-server is
          pinned to one EJB-server all the time.
          If we kill the second servlet-server as from now on round robin works in the
          EJB-server cluster like we expected.
          As we changed this configuration and also use an IPlanet webserver (one
          instance) with the weblogic proxy-plugin which connects to the
          servlet-cluster (2 instances)... round robin is working properly in the
          EJB-server cluster when the client connects to the IPlanet-server....
          what is the problem, why doesn't round robin work in the first
          configuration? any suggestions?
          Torsten Jenkner


Okay, is it that the round-robining is not perfect (i.e., it doesn't send the
          first request to server1 and the second request to server2 and so on) or is it
          sending every request to one server?
          Torsten Jenkner wrote:
          > yes exactly Robert that's it!
          >
          > Torsten
          >
          > "Robert Patrick" <[email protected]> schrieb im Newsbeitrag
          > news:[email protected]...
          > > I don't understand this. Are you saying that with servlet node 1 or
          > servlet
          > > node 2 running by themselves that RR works fine but once you start both
          > > together, it does not?
          > >
          > > This sounds like a case for support...
          > > Robert
          > >
          > > Torsten Jenkner wrote:
          > >
          > > > > > it is always the same stub in our case. Also the round robin in the
          > EJB
          > > > > > cluster does with the same java classes but does not if the servlet
          > > > cluster
          > > > > > consits of two nodes. So it seems to be a configuration problem, but
          > > > which
          > > > > > one?
          > > > >
          > > > > Hmm.... I don't quite understand this last statement about the EJB
          > and
          > > > servlet
          > > > > cluster, could you explain a little more?
          > > > >
          > > > > Thanks,
          > > > > Robert
          > > > >
          > > >
          > > > if you use 2 servlet cluster nodes round robin between the
          > servlet-cluster
          > > > and the ejb-cluster doesn't work. if i kill the second servlet-node it
          > works
          > > > fine. so my thought was it has to be a configuration problem, because RR
          > > > does with one servlet-node. ok?
          > > >
          > > > Torsten Jenkner
          > >

Apache Proxy Plugin with SSL in Weblogic Cluster

Hi,
I have configured a weblogic cluster and configured SSL. Then I configured the apache plugin to work with the cluster machines with non ssl and worked succesfull but when I configured the ssl communication between apache and weblogic I´m having problems.
The actual configuration is:
<Location /spmlws>
SetHandler weblogic-handler
WLLogFile /var/log/httpd/tmpweblogic1.log
DebugConfigInfo ON
Debug ALL
KeepAliveEnabled ON
KeepAliveSecs 15
WebLogicPort 7002
SecureProxy ON
TrustedCAFile /opt/freeware/etc/httpd/conf/trustedCA35cert.pem
TrustedCAFile /opt/freeware/etc/httpd/conf/trustedCA36cert.pem
WLProxySSL ON
RequireSSLHostMatch false
WebLogicCluster machine35:7002,machine36:7002
EnforceBasicConstraints false
</Location>
The problem is that the plugin always takes the last TrustedCAFile. In this way if machine36 is down the plugin tries to send all the request to machine35 but it takes the TrustedCAFile for the machine36 (/opt/freeware/etc/httpd/conf/trustedCA36cert.pem) hence the apache complains
[Wed Jun 30 11:13:56 2010] [error] [client 10.19.232.249] ap_proxy: trying GET /spmlws/OIMProvisioning at backend host '10.19.232.97/7002; got exception 'WRITE_ERROR_TO_SERVER [os error=0, line 796 of ../nsapi/URL.cpp]: '
What can I do to have multiple TrustedCAFile or to have working the communication between apache and weblogic cluster using SSL?
thanks in advance

Acording to the documentation this is not possible.
One way to achieve the load balancing of n-weblogic servers in cluster using ssl is to configure de HttpClusterServlet.

Singleton Classes in Weblogic cluster

Hi
Our application is having singleton classes which we refresh programatically ;though very rarely. We need to move our application into a weblogic cluster. The sigleton class refresh also needs to reflect across the servers :
I could see the SingletonService API here http://download.oracle.com/docs/cd/E12839_01/apirefs.1111/e13941/weblogic/cluster/singleton/SingletonService.html. This contain an activate() and deactivate() .Can we call this activate() method to refresh the singleton object whenever there is a refresh required ? Can we define multiple singleton service classes in the cluster ?
Thanks
Suneesh

I've same issue with singleton service...
1) I’ve created a cluster MyCluster with Man1 & Man2 managed instances.
2) I’ve created singleton.jar out of TestCache.java, which is a singleton cache and have deployed it to the MyCluster.
3) I’ve created man.jar out of StartupClassMan.java, which is a startup class and have deployed it to the MyCluster. And I’ve specified this as startup class for the whole MyCluster.
4) For MyCluster, I’ve specified com.mytest.TestCache.class as singleton service with preferred server as Man1 with migratable to Man2.
5) I run Man1, startup class StartupClassMan is invoked and it gets reference to TestCache, populates the cache & prints size as 1.
6) Now when I run Man2, startup class StartupClassMan is invoked and it get references to TestCache & populates the cache & prints size as 1 only.
For some reason, TestCache is not considered as true singleton. If it is true singleton, running Man2 should have printed size 2 as Man1 has already put one entry into the cache.
And I also don't see activate & deactivate method being called when I run Man1 or Man2 servers. Any help on this is really appreciated.
package com.mytest;
import java.util.Map;
import java.util.HashMap;
public class TestCache implements weblogic.cluster.singleton.SingletonService
private static TestCache testCache = new TestCache();
private Map<String, String> mapCache = new HashMap<String, String>();
private TestCache() {}
public void activate() {
System.out.println("TestCache activate called");
public void deactivate() {
System.out.println("TestCache deactivate called");
public static TestCache getTestCache() {
return testCache;
public void put(String key, String value) {
mapCache.put(key, value);
public String get(String key) {
return mapCache.get(key);
public void remove(String key) {
mapCache.remove(key);
public int size() {
return mapCache.size();
package com.mytest;
public class StartupClassMan
public static void main(String args[]) {
System.out.println("StartupClassMan loaded");
TestCache testCache = TestCache.getTestCache();
testCache.put("Man1","Man1");
System.out.println("size: "+testCache.size());
}

Issue listeneing queue from Weblogic Cluster server with multiple managed server

Haveing issue listeneing queue from Weblogic Cluster server with multiple managed server.
Weblogic Cluster structure is like
Weblogic Cluster01
--ManagedServer01(http://server01.myhost.com:7001)
--ManagedServer02(http://server02.myhost.com:7001)
JMS Servers
JMSserver01 targeting: ManagedServer01
JMSserver02 targeting: ManagedServer02
JMSmodule
ConnectionFactory01 targeting:JMSserver01,JMSserver02
UDQueue01 targeting:JMSserver01,JMSserver02
Uniform Distributed Queue in Monitoring tab showing like this
mysystemmodule!JMSserver01@UDQueue01
mysystemmodule!JMSserver02@UDQueue01
So when I am sending message to any Host(by specifying the provider URL) its distributing equally on both server like
mysystemmodule!JMSserver01@UDQueue01 10
mysystemmodule!JMSserver02@UDQueue01 10
But when try to listen message from these queue, it is listening from one server, for which URL given to connect.
mysystemmodule!JMSserver01@UDQueue01 0
mysystemmodule!JMSserver02@UDQueue01 10
untill I connect to other server by giveing its URL, will not able to access other message left on the queue.
Solutions that tried
1) we have tried give both server URL coma sparated in provider URL
we need to configur same scenario for 5 managed server with 3 listener on other servers.
Do any one have solution for this.

You need to have:
1. Consumers connected to each UDQ member
OR
2. If no consumers in some of the members is expected, you can configure Forward Delay (specify the amount of time, in seconds, that a queue member with messages, but with no consumers, will wait before forwarding its messages to other queue members that do have consumers):
http://docs.oracle.com/cd/E12839_01/apirefs.1111/e13952/taskhelp/jms_modules/distributed_queues/ConfigureUDQGeneral.html
For example you can set it to 10 (10s)
Additional Information here:
http://docs.oracle.com/cd/E23943_01/web.1111/e13727/dds.htm#i1314228
http://docs.oracle.com/cd/E23943_01/apirefs.1111/e13951/mbeans/DistributedQueueBean.html?skipReload=true#ForwardDelay
How Does JMS Load Balancing Work with Distributed Queues and Uniform Distributed Queues? (Doc ID 827294.1)
I hope this helps
Best Regards
Luz

Weblogic cluster for 24x7 environment and a front OHS server

Hi experts,
We are going to set up a reliable J2EE application server environment by using clustering webgoic servers on TWO nodes, and a front OHS https server in DMZ for load balancing.
I am new to weblogic cluster and load balance by OHS. I have visited this forum for Fusion MiddleWare clustering etc. Can some experts share some light on this?
Will our deployment architecture be able to handle the J2EE applications failover? In other word, we can restart one of Weblogic Managed Servers when a new release of J2EE codes are re-deployed without impacting business end users’s usage?
Problems we want to solve:
1. All J2EE applications are available for 24x7 even when new J2EE codes are released and deployed on weblogic Managed Servers ANY TIME, side-by-side deployment and restart a managed server if we need to clean the HTTP cache.
2. The J2EE applications should be accessed by external and internal users with a Single Access Point, like
https://apps.company1.com/j2ee1
https://apps.company2.com/j2ee1
https://apps.company1.com/j2ee2
https://apps.company2.com/j2ee2
All J2EE applications (j2ee1, j2ee2, j2ee3 etc) should be deployed on both of weblogic Servers in a cluster, and pointing to a SAME backend database.
Can some experts share with us the best practices on components and configurations? Thanks.

Seems your architecture is like Browser => OHS ( DMZ ) => Weblogic => DB OR Browser => HLB( like bigip ) => OHS ( DMZ ) => Weblogic => DB
Cluster is the solution for load balancing however if you are using OHS for redirection to weblogic then OHS does the load balancing in round robin way. using cluster in this way has a benefit of in case of any one of your managed server is down the OHS will divert connection request to any one of the active managed server ( you have to turn on dynamic list on at OHS ).
http://weblogicserveradministration.blogspot.com/2010/10/load-balancing-in-weblogic-server.html
Failover is something different, if in case any of the any managed server goes down then your user you get the application session from another server but new one, means the the tasks not saved by the users on earlier session will lost. for that yo need to use cluster and then need to enable the session replication. another best option is you can use the coherence web if you are using latest versions of weblogic supporting coherence web. with that you not need to worry on user sessions and you can start any of the managed server anytime without worrying about the user sessions.
http://weblogicserveradministration.blogspot.com/2010/10/manage-http-session-states-session.html
http://weblogicserveradministration.blogspot.com/2011/05/oracle-coherence-37-coherenceweb.html
http://weblogicserveradministration.blogspot.com/2010/11/clustering-part-i.html
another way is, you can use side by side deployment feature in case you don't want shutdown your application completely, with this, old connections and new requests will goes to old application and once new application activated all new requests will come to new application and once all requests on old application will complete that application will retire automatically.
Regards
Mukesh Negi
http://weblogicserveradministration.blogspot.com

StreamCorruptedException when starting Weblogic cluster

Hi, I have an application running on Weblogic Server 8.1 sp4 in clustered environment. When the cluster is restarted the first server restarts without errors but the second one has the following error message. What does the error message mean? Some messages have disappeared form the JMS queue when the cluster has been restarted. Does this error message have something to do with the disappeared messages?
          <29.9.2006 14:09:17 EEST> <Notice> <Cluster> <BEA-000138> <Listening for announcements from cluster Cluster2 on 237.0.0.6:10603.>
          <29.9.2006 14:09:17 EEST> <Notice> <Cluster> <BEA-000133> <Waiting to synchronize with other running members of Cluster2.>
          <29.9.2006 14:09:47 EEST> <Notice> <Cluster> <BEA-000142> <Trying to download cluster JNDI tree from server CL2Server2.>
          <29.9.2006 14:09:54 EEST> <Error> <JMS> <BEA-040368> <The following exception has occurred:
          java.io.StreamCorruptedException: Unknown object stream version. 5505025
               at weblogic.jms.store.BufferDataInputStream.readObject(BufferDataInputStream.java:167)
               at weblogic.jms.store.JDBCIOStream.doRecoverBodies(JDBCIOStream.java:1024)
               at weblogic.jms.store.JDBCIOStream.doRecover(JDBCIOStream.java:939)
               at weblogic.jms.store.JDBCIOStream.recover(JDBCIOStream.java:1140)
               at weblogic.jms.store.JMSStore.recover(JMSStore.java:315)
               at weblogic.jms.backend.BEStore.open(BEStore.java:264)
               at weblogic.jms.backend.BEStore.start(BEStore.java:151)
               at weblogic.jms.backend.BackEnd.openStores(BackEnd.java:1171)
               at weblogic.jms.backend.BackEnd.resume(BackEnd.java:1290)
               at weblogic.jms.backend.BackEnd.migratableActivate(BackEnd.java:2939)
               at weblogic.cluster.migration.MigratableGroup.add(MigratableGroup.java:107)
               at weblogic.cluster.migration.MigrationManager.privateRegister(MigrationManager.java:180)
               at weblogic.cluster.migration.MigrationManager.register(MigrationManager.java:127)
               at weblogic.jms.JMSService.addJMSServer(JMSService.java:2226)
               at weblogic.jms.JMSService.addDeployment(JMSService.java:2031)
               at weblogic.management.mbeans.custom.DeploymentTarget.addDeployment(DeploymentTarget.java:337)
               at weblogic.management.mbeans.custom.DeploymentTarget.addDeployments(DeploymentTarget.java:597)
               at weblogic.management.mbeans.custom.DeploymentTarget.updateServerDeployments(DeploymentTarget.java:575)
               at weblogic.management.mbeans.custom.DeploymentTarget.updateDeployments(DeploymentTarget.java:241)
               at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
               at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
               at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
               at java.lang.reflect.Method.invoke(Method.java:324)
               at weblogic.management.internal.DynamicMBeanImpl.invokeLocally(DynamicMBeanImpl.java:754)
               at weblogic.management.internal.DynamicMBeanImpl.invoke(DynamicMBeanImpl.java:733)
               at weblogic.management.internal.ConfigurationMBeanImpl.invoke(ConfigurationMBeanImpl.java:509)
               at com.sun.management.jmx.MBeanServerImpl.invoke(MBeanServerImpl.java:1560)
               at com.sun.management.jmx.MBeanServerImpl.invoke(MBeanServerImpl.java:1528)
               at weblogic.management.internal.RemoteMBeanServerImpl.private_invoke(RemoteMBeanServerImpl.java:988)
               at weblogic.management.internal.RemoteMBeanServerImpl.invoke(RemoteMBeanServerImpl.java:946)
               at weblogic.management.internal.MBeanProxy.invoke(MBeanProxy.java:954)
               at weblogic.management.internal.MBeanProxy.invokeForCachingStub(MBeanProxy.java:481)
               at weblogic.management.configuration.ServerMBean_Stub.updateDeployments(ServerMBean_Stub.java:7691)
               at weblogic.management.deploy.slave.SlaveDeployer.updateServerDeployments(SlaveDeployer.java:1304)
               at weblogic.management.deploy.slave.SlaveDeployer.resume(SlaveDeployer.java:347)
               at weblogic.management.deploy.DeploymentManagerServerLifeCycleImpl.resume(DeploymentManagerServerLifeCycleImpl.java:229)
               at weblogic.t3.srvr.SubsystemManager.resume(SubsystemManager.java:131)
               at weblogic.t3.srvr.T3Srvr.resume(T3Srvr.java:966)
               at weblogic.t3.srvr.T3Srvr.run(T3Srvr.java:361)
               at weblogic.Server.main(Server.java:32)

Make sure that the JMS server isn't attempting to load a store that was created from a later version of WLS. Also, make sure that no two JMS servers share the same database table - otherwise they will corrupt each-other's data. I think the latter is likely the problem.
          Tom

Error when deploy to weblogic cluster

My cluster has 2 servers A, B. when I deploy an WebCenter Custom Portal application to weblogic cluster, application on server A work well and application on server B is failed. Log on failed server is “what do i do: seems an odd quirk of the EJB spec. The exception is:java.lang.NoClassDefFoundError: Oracle/jdeveloper/db/adapter/DatabaseProvider”
When I reconfig cluster and then deploy, application on server B work well but application on server A is failed.
Please advise what should I do in this situation?

Hi,
It seems your Application is running with non-seralizable objects where non-searlizable are not supported on cluster environment.
Please check server logs whether you have such warning messages are not.
Let me know if you need any further help on this issue.
Regards,
Kal

Admin Server Funcationality in a WebLogic Cluster

Is this a correct statement of how the Admin Server is handled in WebLogic Cluster:
"If you have a two node WebLogic 10.3 Cluster then you will have the Admin Server for the domain running on node1. The managed server within the domain will run on both node1 and node2. The managed server services your application(s).
If node1 fails then the managed server running on node2 will continue to operate, thus meeting your high availability requirement.
If you need to administer your WebLogic installation you will first have to restore node1 to an operational state so that you can access the Admin Server."
Thank you

Answers :-
Q>"If you have a two node WebLogic 10.3 Cluster then you will have the Admin Server for the domain running on node1. The managed server within the domain will run on both node1 and node2. The managed server services your application(s).
A>Admin server can be running on any node(Physical machine). One Managed server(say MS1) can run only on one node. Two nodes can host two different managed servers. Yes, managed servers services your applications.
Q>If node1 fails then the managed server running on node2 will continue to operate, thus meeting your high availability requirement.
A> Yes, that is the reason of using clustering : Load balancing and failover.
Q>If you need to administer your WebLogic installation you will first have to restore node1 to an operational state so that you can access the Admin Server."
Yes, if admin server was running on node1, then it needs to be up and running to administer the domain.

Singleton class in WebLogic Cluster

Hi,
We have an application set-up in a weblogic cluster.
We have a singleton class in the application.
Since we have two managed servers in the cluster the singleton has two instances, one in each server.
So the basic purpose of use of singleton is lost.
Could anyone please give some alternative implementation.
Regards
Gunajit K

One alternative is to set up some type of centralized store for the information you are trying to share between VMs. I would look at what you are trying to do with the singleton and make some descisions.
Specifically, I would at least look at how much access you will need to the data, how much network traffic it will take to send it to each node in your cluster, and what kind of transactional integrity you require.
A JMS publish/subscribe topic would be a good solution if you are caching some type of data that changes relatively infrequently. You could set this up yourself or check out SpiritSoft's Spirit Cache product (no, I don't work for SpiritSoft).
Or you could simply store the information in some type of database and access it via EJBs or JDBC or whatever you like.
Clearly, the performance of such a store will not be as quick as the in-memory access a Singleton provides inside a single JVM. So be careful about what solution you choose and be sure to optimize its performance as best you can.

Singletons in WebLogic cluster

In my clustered EJB application, there are 3 stateless session beans that need to run as singleton services on exactly one node. So I implemented the interface weblogic.cluster.singleton.SingletonService in these beans with the activate and deactivate methods. The beans and all classes of the application are contained in an ear file I deploy to the entire cluster. Then I configured the bean classes as singletons using the WebLogic console administration (cluster details -> singleton service). After I activated this in changes center, I see a ClassNotFoundException with the class name of one of the beans. So I extracted the ear file to the domain/lib directory, restarted the server and tried again to configure the singleton classes. This time the activation failed because of another reason (HibernateException). I think the reason is, this operation is done in the context of my administration server, not of the cluster server. It seems this doesn't work.
Is it possible to deploy the EJB application with the beans as singletons?
I already found another solution: to set the maximum number of beans in weblogic-ejb-jar.xml, field max-beans-in-free-pool, to 1. This solution is a little bit dirty, after starting the servers, I get an exception on all other nodes complaining that it is not possible to deploy this bean. What I don't know: how would it be handled, if the node with the deployed beans is shut down? Are the beans deployed on the other nodes automatically (as I expect for singleton services) or keep undeployed?
Thanks in advance!
BTW: I'm using WebLogic 10.3 on Windows.

Maybe this will help: http://middlewaremagic.com/weblogic/?p=5799
Also to make Hibernate run on WebLogic you have perform certain steps, there is
problem with the antlr version that ships with Hibernate as another version of
antlr is already loaded in the system classpath of WebLogic.
Some instructions are presented here: http://middlewaremagic.com/weblogic/?p=5861

Sizing of WebLogic Cluster Servers

Similar Messages

Maybe you are looking for