2 vs 1 instance of WLS ?

Greetings !!!
Leaving other parametes same, would 2 instances of WLS with smaller
values of Heap size, executeThreadCount, DBConnections etc. each would
provide better perfrormance than one instances of WLS with higher number
for these paramaters ? Has anyone done benchmarks for such a scenario???
Environement: 2-CPU box, 1GB RAM, Solaris 6, JDK_1.2.2_05a & WLS
5.1-SP5
Thanks for your help/suggestions/pointers???
Sam

Mike,
Originally, the 'more, smaller instances' was a work-around for poorgarbage
collection. I believe that the later JDK's are much better at it. Wouldn'tit
make more sense to simply fix the garbage collection?There is the GC issue, but it is mostly a non-issue now with Hotspot. IBM's
JDKs also have much-improved GC implementations.
It's a queuing theory thing - 1 pool of 50 db connections will be able tostand
more load than 2 pools of 50 db connections, or 10 pools of 5 dbconnections.
Same goes for execution threads.Au contraire ... the front queue (e.g. a hardware load balancer) may
subscribe to that theory, but consider the bank vs. the check-out line
example from CS 101.
However, I don't believe that is the real issue. It's all about how many
requests you complete per second per unit of hardware. Obviously if one
instance works the best, then you stick with that.
Oh, I forgot one - when you start up a cluster with many instances, yournetwork
traffic goes wild. I think this was fixed around 4.5.1 SP 10.Are you telling me that there was a release of WL before 5.1? ;-)
And they have 'director' in front of their names.See below.
Peace,
Director Cameron Purdy
Tangosol, Inc.
http://www.tangosol.com
+1.617.623.5782
WebLogic Consulting Available
"Mike Reiche" <[email protected]> wrote in message
news:[email protected]...
>
"Cameron Purdy" <[email protected]> wrote:
Hi Mike,
Someone needs to send this to www.urbanlegends.com.I'd like to do some testing with WL6.0/JDK1.3/Hotspot 2.0/Solaris 8 etc.,
and in that case you might be right. On some systems, you can crank the
execute threads way up and it won't make a difference -- the system will
not
get saturated.Originally, the 'more, smaller instances' was a work-around for poorgarbage
collection. I believe that the later JDK's are much better at it. Wouldn'tit
make more sense to simply fix the garbage collection?
Let's see what else - you also have to cut your connection pool sizes
in
half.
Most apps keep the pool size about the same as the exec thread count.It's a queuing theory thing - 1 pool of 50 db connections will be able tostand
more load than 2 pools of 50 db connections, or 10 pools of 5 dbconnections.
Same goes for execution threads.
Oh, I forgot one - when you start up a cluster with many instances, yournetwork
traffic goes wild. I think this was fixed around 4.5.1 SP 10.
We run two instances per box here and I wish we didn't.Based on your stated feelings, is there any reason why you run two
instances
per box?Because there are people here that read about what what a great
idea it was. And they have 'director' in front of their names.
Cameron Purdy
Tangosol, Inc.
http://www.tangosol.com
+1.617.623.5782
WebLogic Consulting Available
"Mike Reiche" <[email protected]> wrote in message
news:[email protected]...
Someone needs to send this to www.urbanlegends.com.
Using two instances vs. one instance doubles all kinds of things -
memory
required, overhead,
context switching, all class loading and JSP compiling, configuration
and
maintenance. And if you
use the same jsp workingDir for both WLS the two WLS will overwrite
each
other's generated
class files. (and they don't warn you about that in the doc). Let's seewhat else - you also have
to cut your connection pool sizes in half.
If you're not getting enough concurrency, then increase the number ofexecuteThreads.
If you want reliability, then fix whatever is breaking. It always
better
to not fail than to
recover nicely from a failure.
We run two instances per box here and I wish we didn't.
Mike
"Cameron Purdy" <[email protected]> wrote:
Sam,
It is best to test with your particular application but ... in most
cases,
adding a second instance of WebLogic to a particular host will make
more
effective use of the processor(s) in the server. Since you are payingper
processor, why not take full advantage of them? ;-)
Also, using two network interfaces, you can bind one WL instance to
one
NIC
and the other WL instance to the other NIC and thus not lose the site
if
one
of the networks goes down. (I don't know yet how to force WL to
assign
the
secondary from the other network, though.)
The main variable that you are looking for is throughput: Can you
handle
more concurrent requests? This measurement is basically the number of
typical requests per period of time that are possible before you seethe
server queues growing. In other words, once the server queues start
growing, that means that your cluster isn't managing to keep up, andyou
have lost the war.
Peace,
Cameron Purdy
Tangosol, Inc.
http://www.tangosol.com
+1.617.623.5782
WebLogic Consulting Available
"sam ernie" <[email protected]> wrote in message
news:[email protected]...
Greetings !!!
Leaving other parametes same, would 2 instances of WLS with smaller
values of Heap size, executeThreadCount, DBConnections etc. each
would
provide better perfrormance than one instances of WLS with highernumber
for these paramaters ? Has anyone done benchmarks for such ascenario???
Environement: 2-CPU box, 1GB RAM, Solaris 6, JDK_1.2.2_05a & WLS
5.1-SP5
Thanks for your help/suggestions/pointers???
Sam

Similar Messages

  • Splitting traffic between 2 diff instances of WLS (one  5.1 & other 7.0)

    Hi,
    At my company, we are exploring the possibility of upgrading the appserver for
    our website from WLS 5.1 to WLS 7.0. We run iPlanet web server. However, to reduce
    risk and short-term costs, we are looking at only migrating a few key web-pages
    to WLS 7.0 and keeping the remaining pages on the existing instance of WLS 5.1.
    We want to maintain a single instance of the iPlanet web server.
    I want to know whether this is a viable strategy i.e. splitting the web-traffic
    to 2 different instances of WLS running different versions.
    Thanks for your help.

    You cannot register 2 versions of our plugin in one iplanet instance
    b'coz the function names are same. So in reality it's not possible to
    have 1 iplanet instance to route requests to 2 different WLS versions.
    Kumar
    Andy Davis wrote:
    Hi,
    At my company, we are exploring the possibility of upgrading the appserver for
    our website from WLS 5.1 to WLS 7.0. We run iPlanet web server. However, to reduce
    risk and short-term costs, we are looking at only migrating a few key web-pages
    to WLS 7.0 and keeping the remaining pages on the existing instance of WLS 5.1.
    We want to maintain a single instance of the iPlanet web server.
    I want to know whether this is a viable strategy i.e. splitting the web-traffic
    to 2 different instances of WLS running different versions.
    Thanks for your help.

  • Deploying multiple apps on one instance of wls 8.1

    We currently have have one huge application (one ear file) running on one instance
    of wls 8.1. We plan to break this up in 3 different application (3 different ear
    files). We are using split
    development directory structure of weblogic 8.1. Out problem is, we have some
    ejbs and some war files (like servlets anf jsps) which are common across all the
    applications. How do we manage this ? Do we make a different application for this
    common component ?
    Thanks in advance.
    Yogesh

    Out of curiousity, why are you splitting up the ear?
    How do you intend to package the common ejbs/webapps across the 3 ear files?
    If you're going to copy the common files into each application, then i
    would probably have a separate build.xml that builds the common pieces.
    The other apps can just copy these common pieces into their build
    directory. If the common pieces aren't changing very frequently, then
    this is pretty nice.
    -- Rob
    Yogesh Ranjan wrote:
    We currently have have one huge application (one ear file) running on one instance
    of wls 8.1. We plan to break this up in 3 different application (3 different ear
    files). We are using split
    development directory structure of weblogic 8.1. Out problem is, we have some
    ejbs and some war files (like servlets anf jsps) which are common across all the
    applications. How do we manage this ? Do we make a different application for this
    common component ?
    Thanks in advance.
    Yogesh

  • Multiple instance of WLS in one and same fusion home.

    Hi
    Some SOA/weblogic server related questions
    1. Is it possible to have different instances of Weblogic Server (different versions/patches) installed in one SOA environment, sharing same MW_HOME?
    2. Will there be any problem when patching any of the Weblogic server instances?
    Thanks
    Regards
    /Martin
    Edited by: 890736 on 2011-okt-17 05:15

    Hi Martin
    1. NO, you cannot have different instances of WLS with same version of SOA and using same middleware home. The underlying Weblogic Server version is tightly coupled with the SOA Version deployed on top of it.
    2. We cannot mismatch wls and soa versions, even though they may work. But it is not supported and not recommended. Infact corresponding IDE like JDev version should also match accordingly. And finally the RCU version should also match.
    WLS 10.3.5 -> SOA 11.3.5
    WLS 10.3.4 -> SOA 11.3.4
    WLS 10.3.3 -> SOA 11.3.3
    Thanks
    Ravi Jegga

  • Multiple instances of WLS 5.1

    I have two different applications using WLS. Is it possible to have these applications
    servers on the same box with multiple cpu's. Please note that this is different
    to clustering the same application on one box.
    What are the pros and cons of sharing one box (W2K) for two non integrated app.
    some have recommended that it saves cost of the hardware. I wonder if the cost
    of hardware offsets the dependencies that one incurs having two applications on
    one box (QA, uptime, downtime, scalability..) etc etc.
    Would like to hear some thoughts and maybe white papers..
    Thx

    Not uncommon; works fine. W2K? Not as good as unix for uptime and running
    more than one copy of notepad at a time ;-)
    Peace,
    Cameron Purdy
    Tangosol Inc.
    << Tangosol Server: How Weblogic applications are customized >>
    << Download now from http://www.tangosol.com/download.jsp >>
    "Sunny" <[email protected]> wrote in message
    news:3b7304ab$[email protected]..
    >
    I have two different applications using WLS. Is it possible to have theseapplications
    servers on the same box with multiple cpu's. Please note that this isdifferent
    to clustering the same application on one box.
    What are the pros and cons of sharing one box (W2K) for two non integratedapp.
    some have recommended that it saves cost of the hardware. I wonder if thecost
    of hardware offsets the dependencies that one incurs having twoapplications on
    one box (QA, uptime, downtime, scalability..) etc etc.
    Would like to hear some thoughts and maybe white papers..
    Thx

  • How to identify which process belongs to particular instance of WLS running

    Hi,
    I have a simple question for which i couldn't find an answer. Suppose i have 6 instances of weblogic servers running on a Unix OS. I want to kill a particular instance using "kill" command. I do a "ps -ef | grep java" and find all the instances are point to the same jdk installation in the same directory. For Ex:
    root 17561 17557 0 17:07:32 pts/11 0:00 grep java
    root 6953 1 0 Feb 24 ? 1:35 /usr/java/bin/../bin/../jre/bin/../bin/sparc/native_threads/java -Dviper.fifo.p
    bea81 17258 1 0 Feb 23 ? 0:33 /opt/bea81sp3/jdk142_08/bin/java -client -Xms32m -Xmx200m -XX:MaxPermSize=128m
    bea81 150 1 0 Feb 22 ? 25:20 /opt/bea81sp3/jdk142_08/bin/java -server -XX:NewSize=300m -XX:MaxNewSize=300m -
    bea81 19 4 0 Feb 22 ? 1:29 /opt/bea81sp3/jdk142_08/bin/java -client -Xms32m -Xmx200m -XX:MaxPermSize=128m
    bea81 110 1 0 Feb 22 ? 1:44 /opt/bea81sp3/jdk142_08/bin/java -server -Xms32m -Xmx200m -XX:MaxPermSize=128m
    bea81 25692 25677 0 Feb 17 ? 0:09 /opt/bea81sp3/jdk142_08/bin/java -client -Xms32m -Xmx200m -Xverify:none -Djava.
    How do i identify the right instance i need to kill for a particular domain? Is there any specific unix command?
    Any help will be appreciated
    Megabyte

    mega byet schrieb:
    Hi,
    I have a simple question for which i couldn't find an answer. Suppose i have 6 instances of weblogic servers running on a Unix OS. I want to kill a particular instance using "kill" command. I do a "ps -ef | grep java" and find all the instances are point to the same jdk installation in the same directory. For Ex:
    root 17561 17557 0 17:07:32 pts/11 0:00 grep java
    root 6953 1 0 Feb 24 ? 1:35 /usr/java/bin/../bin/../jre/bin/../bin/sparc/native_threads/java -Dviper.fifo.p
    bea81 17258 1 0 Feb 23 ? 0:33 /opt/bea81sp3/jdk142_08/bin/java -client -Xms32m -Xmx200m -XX:MaxPermSize=128m
    bea81 150 1 0 Feb 22 ? 25:20 /opt/bea81sp3/jdk142_08/bin/java -server -XX:NewSize=300m -XX:MaxNewSize=300m -
    bea81 19 4 0 Feb 22 ? 1:29 /opt/bea81sp3/jdk142_08/bin/java -client -Xms32m -Xmx200m -XX:MaxPermSize=128m
    bea81 110 1 0 Feb 22 ? 1:44 /opt/bea81sp3/jdk142_08/bin/java -server -Xms32m -Xmx200m -XX:MaxPermSize=128m
    bea81 25692 25677 0 Feb 17 ? 0:09 /opt/bea81sp3/jdk142_08/bin/java -client -Xms32m -Xmx200m -Xverify:none -Djava.
    How do i identify the right instance i need to kill for a particular domain? Is there any specific unix command?
    Any help will be appreciated
    MegabyteI think, that You know the port using the instance. With
    "lsof -i TCP:XXXX" You can figure out the PID of the java-prozess.
    Regards Ruedi :-)

  • Multiples Instance of WLS in WIN NT or 2000

              How to run multiple instances of weblogic server in Windows NT or Windows 2000
              

    Checl clustering documentation ..
              "Ganesh" <[email protected]> wrote in message
              news:3b0a76b6$[email protected]..
              >
              > How to run multiple instances of weblogic server in Windows NT or Windows
              2000
              

  • Accessing two different wls instances in Apache

    Hi!
    Does anyone know if it is possible to connect to two different instances of wls
    5.1 sp 6 (same IP, different ports) from an apache front-end under the same <VirtualHost>-instance,
    using mod_wl_ssl.so?

    Thanks for your reply.
    It's not the CuCo being initial but my attribute and the collection wrapper of the value node has only one entry, the empty one which is added in the INIT-method of the context node. So it feels like a second instance of my CuCo as inside the other view I have my collection, attribute etc.
    This is my coding to get the CuCo:
    DATA: lr_cucosearch       TYPE REF TO   zl_iccmp_bp_cucosearch_impl.
    lr_cucosearch ?= get_custom_controller( 'ICCMP_BP_SEARCH/CuCoSearch' ).
    This coding I use in both views. This is wrong? I debugged and it looked like it would search for an instance and then return it instead of creating a second one.
    Am I mistaken?
    Thanks.

  • Several WLS 6.1 instances proxied by one Apache instance

    Group,
    Environment: WLS 6.1, Apache 1.3.20, Solaris 8, UDB 7.2
    Is it possible to proxy several instances of WLS 6.1 through one Apache
    instance on port 80 via the <location> tag in the conf? It seems there is
    no way to specify different weblogic ports for each <location> tag (i.e.
    http://www.somewebserver.com/instance1 proxies to weblogic instance1,
    http://www.somewebserver.com/instance2 proxies to weblogic instance2, etc.)
    Is this possible?
    Thanks for any info,
    Darrin Bolt
    LAT, Inc.

    Group,
    Environment: WLS 6.1, Apache 1.3.20, Solaris 8, UDB 7.2
    Is it possible to proxy several instances of WLS 6.1 through one Apache
    instance on port 80 via the <location> tag in the conf? It seems there is
    no way to specify different weblogic ports for each <location> tag (i.e.
    http://www.somewebserver.com/instance1 proxies to weblogic instance1,
    http://www.somewebserver.com/instance2 proxies to weblogic instance2, etc.)
    Is this possible?
    Thanks for any info,
    Darrin Bolt
    LAT, Inc.

  • 2 WLS Instances :relationship in JNDI

    I have 2 instances of WLS. One is my own instance on NT and the other is an instance
    running on a unix server. Can the NT instance get a initial context for the unix
    instance and
    lookup an ejb in the unix instance's jndi tree ? Do I need to do anything special
    to enable this.
    thanks

    "DK" <[email protected]> wrote in message
    news:[email protected]..
    >
    I have 2 instances of WLS. One is my own instance on NT and the other isan instance
    running on a unix server. Can the NT instance get a initial context forthe unix
    instance and
    lookup an ejb in the unix instance's jndi tree ? Do I need to do anythingspecial
    to enable this.
    thanks
    Yes, you can have the WLS instance on NT access the jndi tree on the unix
    server and access EJB's on teh unix server. You do so by specifying the
    provider_url of the target server when creating your InitialContext.
    EG (from the weblogic docs):
    Hashtable env = new Hashtable(5);
    env.put(Context.INITIAL_CONTEXT_FACTORY,
    "weblogic.jndi.WLInitialContextFactory");
    env.put(Context.PROVIDER_URL,
    "t3://weblogicServer:7001");
    Context ctx = new InitialContext(env);
    Brett

  • Configuration failed 11.1.2. compatibility to wls 10.3

    Hello,
    I developed an application for mobile Web using only Trinidad components under JDeveloper 11.1.2
    I created the ear file and started to deploy it on a Weblogic 10.3 and got some problems during the start of the application:
    <21.10.2011 16:24 Uhr MESZ> <Error> <javax.enterprise.resource.webcontainer.jsf.config> <BEA-000000> <Critical error during deployment:
    com.sun.faces.config.ConfigurationException: CONFIGURATION FAILED! null
         at com.sun.faces.config.ConfigManager.initialize(ConfigManager.java:351)
         at com.sun.faces.config.ConfigureListener.contextInitialized(ConfigureListener.java:222)
         at weblogic.servlet.internal.EventsManager$FireContextListenerAction.run(EventsManager.java:481)
         at weblogic.security.acl.internal.AuthenticatedSubject.doAs(AuthenticatedSubject.java:321)
         at weblogic.security.service.SecurityManager.runAs(SecurityManager.java:120)
         at weblogic.servlet.internal.EventsManager.notifyContextCreatedEvent(EventsManager.java:181)
    ... and a lot more
    I got an ADF Faces application which I developed using JDev 11.1.1.4 running on this server without any complains.
    Any help will be greatly appreciated. Thanks in advance.
    Martin
    Edited by: MartinK on 21.10.2011 07:44

    Hi,
    JDeveloper 11.1.2. requires a patched instance of WLS 10.3.5. Even if using Trinidad, the component set you use in JSf 2.0, which is not what you have on 10.1.3.x
    Frank

  • WLS 8.1 Apache Plug-in is changing HTTP status code 999 to 500

    I'm seeing the weblogic server 8.1 sp4 apache plug-in change the HTTP status code of 999 (set by the application) to 500 Internal Server Error. I'm seeing this with Apache 2.0.48 on aix and with Apache 2.0.58 on windows. I've tried the plug-in from 8.1 SP2, SP4 and SP6 with all yielding the same result. I've tried the SSL plug-in too - same result. I'm told by the maker of this application that they are not seeing this issue with the plug-in for SunOne/NES. What does BEA say about the Apache plug-in changing the http status code? Why is it doing this? Why doesn't the SunOne/NES plug-in do it? Can you provide a plug-in that does not do this?

    Try setting
    Debug ALL
    inside your IfModule. All proxy activity will be written to /tmp/wlproxy.log (on
    unixes). If no info appears in the log, Apache is not invoking the plug-in properly.
    HTH
    "Lukas" <[email protected]> wrote:
    >
    I am trying to set up Apache 2.0.49 proxy plug-in for my WLS 8.1 cluster
    (2 managed
    servers). I followed instructions at http://edocs.bea.com/wls/docs81/plugins/apache.html.
    My test web application runs on
    192.168.10.142:5001/public/testwebapp
    and
    192.168.10.142:6001/public/testwebapp
    I have tested my cluster using another instance on WLS and HttpClusterServlet
    and it works fine.
    I have added following section to my httpd.conf
    LoadModule weblogic_module modules/mod_wl_20.so
    <IfModule mod_weblogic.c>
    WebLogicCluster 192.168.10.142:6001,192.168.10.142:5001
    </IfModule>
    <Location /public/testwebapp>
    SetHandler weblogic-handler
    PathTrim /public/testwebapp
    ErrorPage http://myerrorpage1.mydomain.com
    </Location>
    I restarted my wls cluster and apache as well and when I specify http://192.168.10.142/public/testwebapp
    (apache web server) I receive Error 404--Not Found (even thou I have
    tested my
    http://192.168.10.142:5001/public/testwebapp and http://192.168.10.142:6001/public/testwebapp
    - they work correctly).
    There is nothing in apache’s error_log file.
    There is a record in access_log file reading:
    192.168.10.126 - - [08/Jun/2004:16:23:32 -0700] "GET /public/testwebapp
    HTTP/1.1"
    404 1214
    which gives me an impression that “proxy” – forwarding does not work.
    Where should I look for some errors?
    Thanks,
    Lukas

  • WLS 11 Policy migration during EAR deployment

    We are installing an ADF 11g EAR into our development instance of WLS 10.3.1 and are seeing some strange behavior with the ADF policy migration. During the EAR install from either the WLS console, or using WLST scripts, the ear deployment appears to create entries in our domains system-jazn-data.xml file in the .../config/fmwconfig directory. The policies work fine when testing the application. Users in the various application roles see expected security behavior. If we then start and stop a completely different managed server from the console, the entries that were in the system-jazn-data.xml file disappear, but the our application security continues to work as expected. If we then bounce the server our ear is deployed to, the entries show up again in system-jazn-data.xml.
    Can anyone explain this behavior and verify if this is expected?
    Additional Information: We have followed the enterprise deployment guide when setting up our middleware home directory. We have the following directory structure:
    /opt/oracle/admin/snidomain/aserver
    /opt/oracle/admin/snidomain/mserver
    The aserver directory contains the admin server and the mserver directory contains our managed servers. Each directory contains a config/fmwconfig subdirectory each containing a system-jazn-data.xml file. Each also contains a jps-config.xml file that specifies an XML policy provider pointing at "./system-jazn-data.xml". When we install an EAR that utilizes ADF security, the system-jazn-data.xml file in the mserver directory is updated during deployment, but the one in the aserver directory is not. Each recycle of a managed server appears to replace the contents of the system-jazn-data.xml file in the mserver directory.
    Also, when I bring up Fusion Middleware Control and view the roles/policies of the ear, none are displayed in the UI. If I add a role/policy using Fusion Middleware Control, the new role/policy is placed in the system-jazn-data.xml file in the aserver directory, and the contents of the one in the mserver directory is completely replaced with the same contents as the one in the asever directory, overwriting the ones added during the deployment.

    We are installing an ADF 11g EAR into our development instance of WLS 10.3.1 and are seeing some strange behavior with the ADF policy migration. During the EAR install from either the WLS console, or using WLST scripts, the ear deployment appears to create entries in our domains system-jazn-data.xml file in the .../config/fmwconfig directory. The policies work fine when testing the application. Users in the various application roles see expected security behavior. If we then start and stop a completely different managed server from the console, the entries that were in the system-jazn-data.xml file disappear, but the our application security continues to work as expected. If we then bounce the server our ear is deployed to, the entries show up again in system-jazn-data.xml.
    Can anyone explain this behavior and verify if this is expected?
    Additional Information: We have followed the enterprise deployment guide when setting up our middleware home directory. We have the following directory structure:
    /opt/oracle/admin/snidomain/aserver
    /opt/oracle/admin/snidomain/mserver
    The aserver directory contains the admin server and the mserver directory contains our managed servers. Each directory contains a config/fmwconfig subdirectory each containing a system-jazn-data.xml file. Each also contains a jps-config.xml file that specifies an XML policy provider pointing at "./system-jazn-data.xml". When we install an EAR that utilizes ADF security, the system-jazn-data.xml file in the mserver directory is updated during deployment, but the one in the aserver directory is not. Each recycle of a managed server appears to replace the contents of the system-jazn-data.xml file in the mserver directory.
    Also, when I bring up Fusion Middleware Control and view the roles/policies of the ear, none are displayed in the UI. If I add a role/policy using Fusion Middleware Control, the new role/policy is placed in the system-jazn-data.xml file in the aserver directory, and the contents of the one in the mserver directory is completely replaced with the same contents as the one in the asever directory, overwriting the ones added during the deployment.

  • WLS 10.0 JDBC connection pool shrink not working

    We seem to be having a problem with jdbc connection pools not shrinking. In the connection pool properties in the WLS console we have:
    Initial Capacity 4
    Maximum Capacity 15
    Shrink Frequency 900
    However when I look at the underlying xml config file (see below) the last two values above are not present in the config!
    Any ideas what is going on here?
    thankyou,
    Chris
    <?xml version='1.0' encoding='UTF-8'?>
    <jdbc-data-source xmlns="http://www.bea.com/ns/weblogic/90" xmlns:sec="http://www.bea.com/ns/weblogic/90/security" xmlns
    :xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:wls="http://www.bea.com/ns/weblogic/90/security/wls" xsi:schemaLo
    cation="http://www.bea.com/ns/weblogic/920 http://www.bea.com/ns/weblogic/920.xsd">
    <name>mdmrDataSource</name>
    <jdbc-driver-params>
    <url>jdbc:oracle:thin:@(DESCRIPTION = (ADDRESS = (PROTOCOL = TCP) (HOST = xxxrac01-vip.nzlxxx01.com) (PORT = 1521))
    (ADDRESS = (PROTOCOL = TCP) (HOST = xxxrac02-vip.nzlxxx01.com) (PORT = 1521)) (LOAD_BALANCE = yes) (CONNECT_DATA =(SERVE
    R = DEDICATED) (SERVICE_NAME = xxxrac.nzlami01.com)))</url>
    <driver-name>oracle.jdbc.pool.OracleDataSource</driver-name>
    <properties>
    <property>
    <name>user</name>
    <value>xxx_comms</value>
    </property>
    <property>
    <name>fastConnectionFailoverEnabled</name>
    <value>true</value>
    </property>
    <property>
    <name>implicitCachingEnabled</name>
    <value>true</value>
    </property>
    <property>
    <name>connectionCachingEnabled</name>
    <value>true</value>
    </property>
    <property>
    <name>ONSConfiguration</name>
    <value>nodes=xxxrac1:6251,xxxrac2:6251</value>
    </property>
    </properties>
    <password-encrypted>xxx</password-encrypted>
    </jdbc-driver-params>
    <jdbc-connection-pool-params>
    <initial-capacity>4</initial-capacity>
    <test-connections-on-reserve>true</test-connections-on-reserve>
    <test-table-name>SQL SELECT 1 FROM DUAL</test-table-name>
    </jdbc-connection-pool-params>
    <jdbc-data-source-params>
    <jndi-name>jdbc/XXXX</jndi-name>
    <global-transactions-protocol>None</global-transactions-protocol>
    </jdbc-data-source-params>
    </jdbc-data-source>

    You are right that the XML seems to lask those other settings... Did you
    sae those settings from the console?
    This is what's in one of mine....
    <jdbc-connection-pool-params>
    <initial-capacity>0</initial-capacity>
    <max-capacity>300</max-capacity>
    <shrink-frequency-seconds>900</shrink-frequency-seconds>

  • Apache plugin support for WLS  6.1 is not working for me.

    Following are the additional configurations I used in httpd.conf after doing LoadModule
    & AddModule
    Alias /weblogic "/opt/BEA/wlserver6.1/config/petstore/applications"
    <Location /weblogic>
    SetHandler weblogic-handler
    </Location>
    <IfModule mod_weblogic.c>
    WebLogicHost localhost
    WebLogicPort 7003
    </IfModule>
    I get Permission denied & not even 403/forbidden for Petstore example jsps. I
    am trying
    to proxy it by URL basis not on mime type. I tried doing same thing for one of
    my own domains
    & server. The results were same. Now, If I try to comment out the Alias line from
    my httpd.conf
    file, it gives me 404 error. Can anyone help me out in this regard? I doubt my
    configurations
    in my weblogic's config.xml file.
    Thanks,
    Naveen

    Hi.
    I'm not clear on what you are trying to accomplish with the alias pointing to the absolute
    path of your WLS applications directory. This is not how the plugin works. As you have it
    set up now the plugin looks for URLs with /weblogic in them and proxies those requests to a
    running instance of WLS.
    For example, the plugin will accept the following URL and forward it to WLS:
    http://host:port/weblogic/index.html
    WLS will respond IF you have a webapp named weblogic. From your example it looks like you
    are trying to proxy to the petstore example. The URL to access the petstore example that
    ships with wls is something like: http://localhost:7003/estore/control/main so if you want
    to proxy by path you may want to proxy when the path contains /estore instead of /weblogic.
    Hope this helps.
    Michael
    Naveen wrote:
    Following are the additional configurations I used in httpd.conf after doing LoadModule
    & AddModule
    Alias /weblogic "/opt/BEA/wlserver6.1/config/petstore/applications"
    <Location /weblogic>
    SetHandler weblogic-handler
    </Location>
    <IfModule mod_weblogic.c>
    WebLogicHost localhost
    WebLogicPort 7003
    </IfModule>
    I get Permission denied & not even 403/forbidden for Petstore example jsps. I
    am trying
    to proxy it by URL basis not on mime type. I tried doing same thing for one of
    my own domains
    & server. The results were same. Now, If I try to comment out the Alias line from
    my httpd.conf
    file, it gives me 404 error. Can anyone help me out in this regard? I doubt my
    configurations
    in my weblogic's config.xml file.
    Thanks,
    Naveen--
    Michael Young
    Developer Relations Engineer
    BEA Support
    [att1.html]

Maybe you are looking for

  • Installing of Adobe Reader not possible

    Hello, i upgraded my XP Pro. to vers.2009  SP3. Since then i'm trying desperatly to install Adobe Reader-(downloaded with different browsers IE&Chrome and diff. languages )but everytime,i receive a different "Error Code" 1.Code was regarding a missin

  • Number to Word Convert Function

    Dear All, Plz anybody have function name which can convert NUMBER to WORD. example:  1213 to One Thousand Two Hundred Thirteen thanks, RP

  • Transfr dates from child network to parent network(using subnetwork functio

    Hi All, i have created 2 separate projects one is parent with 1 network activity assigned to root wbs element and other is child project with several network activitites. By using sub network function i have assigned the parent project network activi

  • Automatic start of Siri

    Siri is not automatically starting when getting closer to the ear. This was worked in iOS 7. For your comments please

  • PDF issue while writing to application server

    Hi Guys, I am writing my PDF file into application server but there is an additional line coming up in the file after EOF as below. This additional line creates a new blank page when they take printout from web service. %%EOF# ef0 R0000 nRA