How do managed servers share datasource?

Hi all,
my db is experience very high number of connections from my weblogic instance. I would reckon that the number of session configured in my db is too low. In view of this, i would like to find out the following:
when i configure a particular a data source, i can specify the max and initial connection capacity. Assuming i specify max and initial = 50 connections and i target it to 5 managed servers.
How many jdbc connections are created in total when i start up all my servers? only 50? or 250?
That is to say, is the 1 data source shared ( 50 connections shared by 5 servers) or each server will get an instance of 1 datasource ( 50 connections) making a total of 250 connections created?
Edited by beaweblogic2sin at 01/22/2008 2:20 AM

gerald lee wrote:
Hi all,
my db is experience very high number of connections from my weblogic instance. I would reckon that the number of session configured in my db is too low. In view of this, i would like to find out the following:
when i configure a particular a data source, i can specify the max and initial connection capacity. Assuming i specify max and initial = 50 connections and i target it to 5 managed servers.
How many jdbc connections are created in total when i start up all my servers? only 50? or 250?
That is to say, is the 1 data source shared ( 50 connections shared by 5 servers) or each server will get an instance of 1 datasource ( 50 connections) making a total of 250 connections created?
Edited by beaweblogic2sin at 01/22/2008 2:20 AM50 per server = 250. If you are unsure you need 50 per server,
set the initial to less, and let the pool grow if/when needed.
For a server to bear maximum load, each of it's threads
would need it's own connection, so a 25-thread server
would need 25 connections, max. If your load never needs to
involve every thread, then you may get away with fewer.

Similar Messages

How many managed Servers do you need?

Hi,
One thing that has come up time and time again, is when trying to think about a new project and a new WL domain, is trying to know how many WL Managed Servers do you need?
Does BEA have some sort of metrics which says ok for this amount of input we recommend you having 4 managed servers on 4 single boxes?
And is there think on have 2 managed servers on 1 box or 2 managed servers on two boxes?
Any Help Appreciated.
Kind Regards,
Alistair.

Hi :-D
I don't believe BEA have guide lines on how many Managed servers are required based on Metrics. You could look at some performance evaluations but it's really only relevant if your application acts in the same way.
Cheers,
Doug.

Calculating how many managed servers you need.

Hi,
Does anyone know a BEA webpage were it tells you, on the lines of :
If you have 5000 customer hitting your weblogic cluster you only need 1 managed server. If you have 30,000 you need two managed servers. If you have 500,000 you need three.
Does anyone know of a performance page like this?
Regards,
Alistair.

There is no such performance numbers from BEA. I think you need to rely on capacity planning process to estimate number of server based on your workload and performance requirements.
http://dev2dev.bea.com/pub/a/2002/11/Hazarika.html
http://edocs.bea.com/wls/docs81/capplan/capfact.html#1009720
There are some benchmark results for WebLogic are available at
http://edocs.bea.com/wls/docs81/capplan/capbase.html#1054891
http://www.spec.org/jAppServer2004/results/jAppServer2004.html
-Jayesh

How to Separate Management Servers for Ease of Administration?

Hello,
I am fairly new to SCOM, though have been charged with creating a monitoring solution for a particular group of customers and integrating it as far as possible into our existing corporate SCOM implementation, so any help, tips and so on would be greatly
appreciated.
The customer monitoring requirements will be very different to those of our fairly standard corporate environment - we will need to monitor very different technologies, with some systems sat in different domains or a DMZ. Our central servicedesk will still
need to monitor and action alerts and events for both environments as seamlessly as possible.
At the same time, I would like to separate the environments as the configuration of the management servers for the customer environment will be quite different from our corporate one, and I want to avoid having to make changes to all our corporate servers
every time I need to change something on the customer environment - so things like management pack installs, certificate admin, static DNS config and so on.
How would I go about this please? Would it be better to create a separate management group, with servers dedicated to the customer environment, then connect this to the corporate management group so that the servicedesk can still monitor it, or is there
another way of doing it? Perhaps using resource pools? Time will be a factor here, though ultimately the goal is ease of administration going forward.
Thanks

Using resource pools will not allow you to separate configuration such as MP installs, security, etc. as they are all part of the same management group. From Technet, "A resource pool is a collection of management servers used to distribute
work amongst themselves and take over work from a failed member." If configuration needs to be completely separate from the base configuration perspective, you could use separate connected management groups.
Another option is to separate the configuration within the management group. For example, if you need to monitor SQL computers in your corp environment but not the customer environment, simply create a group containing customer computers and override
the SQL discovery. This applies to other technologies as well. You can even group computers by environment (corp/customer) by using reg key attributes like Kevin Holman describes here: (While my account is being verified
I can't post links but you can search "Creating custom dynamic computer groups based on registry keys on agents"). If you are familiar with MP authoring, you could create a new MP with a new Windows Computer based class instead of extending
the Windows Computer class. From a capabilities perspective, you can certainly monitor corporate, customer, DMZ, and other domains in the same management group. If the domain is untrusted or in the case of the DMZ, workgroups computers, you can
utilize gateway servers with certificate authentication which will be configured for each domain/dmz gateway.
Ease of administration is a tricky concept here...if you utilize separate management groups, administration will have to be handled completely separate in 2 separate consoles. If there are separate SCOM admins per management group, this is
of no issue, but if 1 team/person is managing both, this can be difficult. Alternatively, there will be some up front work to using groups to separate the environments using groups, discoveries, etc.

How to publish OSB project to other managed servers?

How to publish OSB project to other managed servers
Edited by: 991757 on 2013-3-4 下午6:09

Not required to be published.

How to install Window service for Managed Servers of a Cluster

Dear All,
          I am preparing a production setup consist of three machines A,B and C. Machines A is the Admin Server which hosts a cluster consist of two managed nodes. I intend to schedule both these nodes, targeted to two different machines (machines B node-b and machine C host node-c), as startup task on daily basis. I have no Node Manager running on machine A, but on machine B and machine C.
          I am only interested to schedule started task for node-b and node-c only. I need to know how to install the Windows services for the node-b and node-c since there are targeted on remote machines. Starting managed server as Win sevice works in the case of standalone managed servers but not sure when it comes to cluster setup.I appreciate if you can throw some light on this question.
          Thanks
          Kris

Krish,
          >Let say I gracefully shutdown both managed servers while >Admin is running. Do you think both these nodes while be >brought up ? I dont think they will come up.
          If you require an automatic restart when the managed server did not enter a failed state you can use the weblogic.Admin utility to ping the servers and kick off a server startup in the normal way with nodemanger see:
          http://e-docs.bea.com/wls/docs81/admin_ref/cli.html
          http://e-docs.bea.com/wls/docs81/adminguide/confignodemgr.html
          If your servers entered a failed state you can use weblogics self health monitoring.
          >Can we use JMX MBean to start managed servers ?
          Yes.
          As long as the nodemanagers are running you can start the managed servers using scripted commands using the weblogic.Admin class or your own java app using JMX.
          I think you may also be able to use WLST
          http://dev.bea.com/codelibrary/code/wlst.jsp
          hope this helps
          cheers
          Please see the
          Hussein Badakhchani
          www.orbism.com

How nodemanager starts managed servers?

Hello All,
I'm facing strange issue with weblogic server. Currently i'm using WLS 8.1 SP 4 on solaris 9 server.
To start weblogic i have customize the script such that when server goes on reboot it will pick the startweblogic.sh script from /opt/bin and start the weblogic. Where as in startweblogic.sh i have mentioned -
NMLOG=/u001/wls8/bea/user_projects/prodDomain/logs/optbin/nodemanager.log; export WL8NMLOG
ADMINLOG=/u001/wls8/bea/user_projects/prodDomain/logs/optbin/adminserver.log; export WL8ADMINLOG
CLUSTERLOG=/u001/wls8/bea/user_projects/prodDomain/logs/optbin/cluster.log; export WL8CLUSTERLOG
cd /u001/wls8/bea/weblogic81/server/bin/
su weblogic -c "./startNodeManager.sh >> $NMLOG 2>&1" &
sleep 60
cd /u001/wls8/bea/user_projects/prodDomain/
su weblogic -c "./startadminserver.sh >> $ADMINLOG 2>&1" &
sleep 1200
cd /u001/wls8/bea/user_projects/prodDomain/
su weblogic -c "./startcluster.sh >> $CLUSTERLOG 2>&1" &
Though the script works as expected till adminserver is UP/Running, but script doesn't wait for 1200 secs. to start cluster. Somehow nodemanager starts the managed servers immediately after the adminserver is UP.
I have checked the nodemanager.properties file, there is nothing such mentioned. Also, from console i have unchecked the option of AutoStart for both managed servers.
Can anyone tell me how nodemanager starts the managed servers automatically??
Thanks,
HK

can you please tell me what is this .startCluster.sh script.
is it your custom script?
And the idea i have in my mind is that after
cd /u001/wls8/bea/weblogic81/server/bin/
su weblogic -c "./startNodeManager.sh >> $NMLOG 2>&1" &
sleep 60
cd /u001/wls8/bea/user_projects/prodDomain/
su weblogic -c "./startadminserver.sh >> $ADMINLOG 2>&1" &
sleep 1200
you can give the command to start the individual managed servers like:
cd /u001/wls8/bea/user_projects/prodDomain/
su weblogic -c "./startManagedWebLogic.sh <arguments> >> $CLUSTERLOG 2>&1" &
I think this way it will work....

How to start managed servers in parallel?

Hello everyone,
I am developing a wlst script that would start all managed servers on a particular machine (machine_name should be provided) from the admin server side. The script works fine. However, since there were some domains that have 25+ managed servers for a single machine, it takes some time to start all of them in sequence. Is there a way to trigger a start command for each of the managed servers that I need to start just like what the node manager does when i start them through admin console? Something like triggering parallel processes running on background. Here is my code snippet:
                #It focuses only on all managed servers defined within the listen address provided
                if (server_listenaddress == managed_server) and (server_name != admin_name):
                        if (execute_command == "start"):
                                print 'About to start',server_name,'on',server_listenaddress,'...'
                                print
                                try:
                                        start(server_name)
                                        print
                                        state(server_name,'Server')
                                        print
                                        nmDisconnect()
                                except WLSTException:
                                        print 'Server instance',server_name,'could not be started. Skipping this server..'
                                        print
                                        state(server_name,'Server')
                        else:
                                print 'About to stop',server_name,'on',server_listenaddress,'...'
                                print
                                try:
                                        shutdown(server_name,force='true')
                                        print
                                        state(server_name,'Server')
                                        print
                                except WLSTException:
                                        print 'Server instance',server_name,'could not be stopped. Skipping this server..'
                                        print
                                        state(server_name,'Server')
                elif (server_name == admin_name):
                        #It skips to restart the admin server since this script it only to restart managed servers.
                        print 'Server',server_name,'is the admin server. Skipping...'
                        print
                else:
                        print 'Managed server',server_name,'is not configured with',managed_server,'listen address. Skipping this server..'
Thanks in advance for all your help!!
regards,
Rustan

Works perfectly, thanks!!!
In summary, I just modified this line:
start(server_name)
and changed it to:
start(server_name, block='false')
In Admin Console, it is set to false but in WLST, it is defaulted to true so we need to explicitly set it to false.
"Oracle Fusion Middleware Oracle WebLogic Scripting Tool, block is always set to true."

How to start Managed Servers in Cluster in background

I am new to weblogic. I have two managed servers each on two boxes, i need to start these managed servers in background.
Thanks

I got this error while starting Managed Server2 on machine A. Managed Server was started using the same command line with clean start. I am using JDK1.5.0_14
eg., nohup ./startManagedWebLogic.sh Managed-mn2 http://localhost:7001 &
<Dec 12, 2008 2:17:11 PM EST> <Notice> <WebLogicServer> <BEA-000365> <Server state changed to FORCE_SHUTTING_DOWN>
[thread 43 also had an error]#
# An unexpected error has been detected by HotSpot Virtual Machine:
# SIGBUS
Exception (0xa)in thread " at pc=0xff390c50CompilerThread0, pid=21379" , tid=1java.lang.OutO
fMemoryError: requested
34416#
bytes# Java VM: Java HotSpot(TM) Server VM (1.5.0_14-b03 mixed mode)
for # Problematic frame:
Chunk::new# . Out of swap space?
C [libc_psr.so.1+0xc50]
An error report file with more information is saved as hs_err_pid21425.log
Exception in thread "CompilerThread0" java.lang.OutOfMemoryError: requested 34416 bytes for Chunk::new. Out of swap space?
Edited by: user10705269 on Dec 12, 2008 11:42 AM

OSB Managed Servers - How to add one to the OSB Cluster

Hi,
I have an OSB Cluster of two managed servers. There is an Admin Server that controls the two.
On the Admin Server I see that there are about 137 odd Applications and a list of Libraries installed.
However on the Managed Servers I see there is not a single application.
Does Managed Server show only the Custom applications? I feel not. Applications like AQAdapter, MQ Adapter, ALSB Routing, etc. should they all not be present on the Managed Servers?
If I create a new Managed Server that I want to Add to the Cluster there is no option to deploy those applications to the newly created Managed server in one shot. The only time it can get those applications is if the Applications were deployed to the OSB Cluster and not to the managed servers.
Is there a way (apart from extending the domains again and having those applications deployed to the OSB cluster) to have all these applications running on the Managed servers as well?

Does Managed Server show only the Custom applications? I feel not. Applications like AQAdapter, MQ Adapter, ALSB Routing, etc. should they all not be present on the Managed Servers?Your understanding is correct. Managed server should have libraries and adapters targetted to it.
Is there a way (apart from extending the domains again and having those applications deployed to the OSB cluster) to have all these applications running on the Managed servers as well? Yes, manually you have to target each and every application to the cluster but it may not be an effective process. The basic funda of cluster is to make applications independent of managed server, so that if you add/remove a managed server then it does not affect the existing configuration. I think you should follow the standard practice and always create a clustered domain where all applications are targetted to cluster and NOT to the managed server. You may like to refer -
http://download.oracle.com/docs/cd/E17904_01/doc.1111/e15022/config.htm#i1531648
Regards,
Anuj

How to start managed servers simultaneously using WLST?

Hi folks,
I am trying to write a simple script to provide a simple way to bounce my env, I am pretty much done. However, my script now starts (and stops) the managed servers sequentially which is not what I wanted.
I am wondering if there is a way to start (and stop) the managed servers within a domain in parallel as I can definitely do that using Admin Console.
I do have an idea in mind:
Write a simple python script to start/stop a single managed server, then use another shell script to trigger the python script in background mode.
But I think this approach is really ugly, and am wondering if there is a native way to do that in WLST.
Thanks for your repose in advance.

This does not solve my problem, what I am trying to do is to being a an indefinite about of managed servers at the same time, not all the managed servers.
For example, suppose I have 5 clusters in a domain, and there are 3 servers in each cluster. I might want to bring up all the major servers first at the same time,
which mean to bring up one server in each cluster at the same time.
Please help.

Deployment failed: No managed servers to service request: how to recover

I am unable to deploy applications, instead get error:
2013-05-14 05:27:53 CDT: Starting action "Deploy Application"
2013-05-14 05:27:53 CDT: Deploy Application started
2013-05-14 05:27:53 CDT: Deployment failed: No managed servers to service request
2013-05-14 05:27:53 CDT: Application deployment failed.
2013-05-14 05:27:53 CDT: "Deploy Application" complete: status FAILED
The error occurs both from Jdeveloper and UI deployment. Is there a way to recover from this ? This did work earlier today..

I am experiencing the same stuck job on my java cloud instance - can Oracle support reset my job
If I try to start/stop/redeploy or add an application I get the following error
"Information
The operation cannot be invoked because the service is running Java Cloud Service Job 170371. Wait for the job to complete, then re-run the operation."
https://forums.oracle.com/thread/2548962
https://forums.oracle.com/thread/2501942?tstart=30
thank you
/michael

How create one cluster whit managed servers in diferent hosts?

in the same domain

Have you created a domain using the config tool ( which gives you just an admin server and no managed servers )?
Start your admin server & login.
Create a session in the Change Center
Create 1 machine within the admin console for each host: ( Domains -> Environment -> Machines - > New ). Use the actual host name for the value of Machine Name.
So create a machine named "myhost1.yourcompany.com" and a machine named "myhost2.yourcompany.com".
You can also do this within the configuration tool:
http://download.oracle.com/docs/cd/E12840_01/common/docs103/confgwiz/custom.html#wp1071071

How to deploy config files to managed servers

We have several configuration files that are used by our EJB application at
startup. Is there a way to automatically deploy these configuration files
to the managed servers. Right now, we are forced to FTP to files over to
the managed servers as part of the deployment process. It would be nice if
there were an easier way.
Randy Ott

why cant you include them inside your ejbjar as resource files ?
prasen
Nitin Shenoy wrote:
>
Guess no one replied!!
"Randy Ott" <[email protected]> wrote in message
news:3b67022e$[email protected]..
We have several configuration files that are used by our EJB applicationat
startup. Is there a way to automatically deploy these configuration files
to the managed servers. Right now, we are forced to FTP to files over to
the managed servers as part of the deployment process. It would be niceif
there were an easier way.
Randy Ott

Hi, we need to create the test environment from our production for oracle AP Imaging. we have soa,ipm,ucm and capture managed servers in our weblogic. can anyone tell me what is the best way to clone the environment, can I just tar the weblogic file syste

Hi, we need to create the test environment from our production for oracle AP Imaging. we have soa,ipm,ucm and capture managed servers in our weblogic..
Can anyone tell me what is the best way to cloning the application from different environment, the test and production are in different physical server.
Can I just tar the weblogic file system and untar it to the new server and make the necessary changes?
Can anyone share their experiences and how to with me?
Thank in advance.
Katherine

Hi Katherine,
yes and no . You need as well weblogic + soa files as the database schemas (soa_infra, mds...).
Please refer to the AMIS Blog: https://technology.amis.nl/2011/08/11/clone-your-oracle-fmw-soa-suite-11g/
HTH
Borys

How do managed servers share datasource?

Similar Messages

Maybe you are looking for