How to config RAM of two instance sap on single server ???

I have a server (HDD: 146 Gb, 18 Gb RAM, RHEL 5.4 64bit, Oracle 10g R2), I used as system DEV and I have two questions:
Question 1: I want to install two instance SAP (SAP ECC 6.0 and BI 7.0) on this same server. I want to spend 9 Gb RAM for each SAP instance. So I have to configure how (Oracle instance is how much each instance SAP) ?
Question 2: Sap Application generally takes much RAM for each instance
Question 3: I need to install the package to install the SAP instance on RHEL 5.4 64 bit?
Question 4: What j2sdk package for 64-bit REHL 5.4 64bit?
thank for your advice
Edited by: myfriend280985 on Dec 31, 2009 6:20 PM

Hi,
Please refer to SAP Notes
1172419 - Linux Supported Java versions on the x86_64 platform
1048303 - Red Hat Enterprise Linux 5.xInstallation and upgrade
171356 SAP software on Linux: Essential information
There is enough info available for all your questions and prepare a plan, start the installation.
Best of Luck.
Cheers.....,
Raghu

Similar Messages

How Can I Use Multiple Weblogic Instances in a Single OS

Hello Everyone,
Actually I have to install Some different applications. Few of them need weblogic 10.3.6 and others need 10.3.4. The OS am using is Oracle Enterprise Linux 5.
Now I am able to install 2 separate(One of 10.3.4 and 10.3.6) instances with two different users,In two different directories.
I have installed the weblogic 10.3.6 version with a user webadmin and installed node manager with port 5556. This is working fine.
The main problem here is :
In the second instance (10.3.4 ) installed with a a different user and gave the port number to NodeManager as 1600 and its not getting started. Its throwing error and also after some errors in the terminal am able to see that its reverting to port number 5556 only.
What might be the issue?
I have to install 2 different versions of weblogic in a single Server. But am failing with NodeManager. What Can I do to have multiple weblogic instances with multiple versions in a single server ?
Can anyone suggest a resolution for this please ?
Thanks in advance.

Pl do not spam these forums with multiple posts - How Can I Use Multiple Weblogic Instances in a Single OS

How to get space between two values in a single cell of a table.

hello,
how to get space between two values in a single cell of a table.
thanks a lot.
kailash.

sorry i got the answer.

Two Instances of Oracle Application Server 10.1.3.x using the same DB

I can't seem to find any documentation about installing two instances of Oracle Application Server 10.1.3.x using the same database.
Would the two instances conflict? Should I create a separate database for each instance? I am attempting to cluster two application servers and I want to know if I will need multiple databases, I can't see a reason why this would be a problem.
Any advice on this matter would be appreciated.

You have probably downloaded the wrong version.
Do not download the Oracle SOA Suite 10g (10.1.3.5.1) for WebLogic Server (10.3.1),
but the one below that option: Oracle SOA Suite 10g (10.1.3.x)
See if you get something like the following zip file: soa_windows_x86_101310_disk1.zip
This is just the OC4J without the WebLogic Server.

How to know login history/events of Sap user from server

how to know login history/events of Sap user from server without operating users computer.
Am using B1 2007A PL47 Forthshift(FSE version 8.5 SP03PL05)

Hi,
Check the thread
Re: User login history SAP Business One
*Close the thread if issue solved.
Regards
Jambulingam.P

Two instances on the same server

I've already created one instance on 11g R2 database, but I was told to create another instance on the same server. Can I use the same listener for both instance? Will it cause any problem?
Thanks!
- Johnny

tanww888 wrote:
Thanks for all your responses. I have not dealt with two instances on the same server before. I'll take your advice to start with the same listener for both instances and see if it is causing any issue. My database server is a Oracle Sparc 5120 with plenty of memory and disk space. The second instance that I'm going to create will be much smaller than the first one.
Single or multiple listeners aren't really a factor for performance. Notice that even in Dan's remarks, none of his reasons for multiple listeners had anything to do with performance.
Question, is it easy to create/configure the second listener to point to the second instance? Just in case there is an issue using the same instance.
The listener doesn't "point to" the database - not in the sense you seem to think. It simply watches the ports it is configured to use (1521 by default) for connection requests from clients. When it gets a request it evaluates whether or not it knows about the service (instance) requested, and if it does it spawns a separate server processes to communicate between the client and the db instance, then is out of the way. The listener creates the connection but does not sustain it. You can stop a listener and all existing connection will continue because the listener has no part in maintaining them.
Multiple listeners would be configured by adding LISTENER sections to the listener.ora file. You might find it instructive to realize that if there is no listener.ora, a 'lsnrctl start' will cause a default named listener to start on the default port of 1521. By default, any db instances on that server will register themselves with the default listener on the default port. After this self-registration, the listener will know about the instance if/when it receives any connection requests for it. Let's switch from the telephone analogy to that of a desk clerk at a hotel. He comes to work and is handed a piece of paper that says "We have Tom, Dick, and Harry registered as guests" This would be analogous to the listener learning of instances via the SID_LIST section of the listener.ora file. Then Fred walks up and says "I'm Fred and I'm in room 238", so the clerk makes a note of that. This would be analogous to the instance registering itself to the listener. Either way, then someone walks up and says "Where would I find Tom (or Fred)". The clerks checks his lists and directs the person to the correct room. He doesn't escort the person around to Tom and act as a go-between for their conversation. And hotel management doesn't want to have to pay a separate clerk for each guest.
You might want to read my articles about TNS in general at edstevensdba.wordpress.com. I think they will help you get a clearer understanding.
Thanks,
- Johnny

Installing/Configuring multiple CMS instances in a single server machine

Hello,
We are installing BO enterprise server and we have a requirement which is as follows..
1. We have been given 1 server machine ( actual physical hardware machine )
2. We have been given two database repository instances ( one for Dev , other for QA )
3. We have to configure the 2 environments in the same server.
This means, Having two different CMS instances and pointing them to 2 different database repositories in one hardware server machine.
Any pointers how this can be achieved ?
I found a link where its mentioned that it can be achieved through the concept of SIA
[http://www.businessobjectstips.com/tips/server/business-objects-xi-3-xs-server-intelligence-agent-sia/]
But how do we achieve it?

First of all - this is not a good idea. It's possible to do, but not a good practice.
The purpose of QA environment is to Test what has been developed. If QA and DEV are on the same machine - what's the point?
Then your hardware server has to be big enough to support this.
I might be better to have 2 VM instances...
As to how to go about doing that :
1. It has been discussed here before, so if you search you'll see some good threads on the subject.
2. You'll need to review Admin guide and Install and Deployment Planing guides for the version of BO software you're going to use. (http://help.sap.com)
3. While pure BO stuff is fairly easy - you just create 2 SIA each pointing to its own DB and use different port ranges for both - you'll need to think about how to go about maintaining 2 instances of the Application server (Tomcat, WAS, WL etc ) to host 2 sets of BOE web apps (Infoview/CMC/dswsbobje)).

I have a lion server set up to host storage for time machine backups. Is there a limit to how many different computers can back up to the single server to back up. I am only backing up the users folders . The backup drive is an external Drobo with 6 TB.

I have a lion server set up to host storage for time machine backups. Is there a limit to how many different computers can back up to the single server to back up. I am only backing up the users folders . The backup drive is an external Drobo with 6 TB. Right now it seems to back up my users all day long. I set their time machine interval to every 5 hours but it still runs all day long and is very slow. Server is on a new Mac Server mini.

AFAIK, there is no user limit.. it is simply a question of the network load. AFP is a fairly clean protocol.. but TM is not.. it does a lot of testing of the backup.. infact your setting of 5hour gap could well make it worse.
I would sort out the clients that are doing 100GB backups to a different backup location.. no matter how fast your network.. 100GB takes time to backup.. and over wireless forever.. it will never end.
I guess some of these clients are on wireless?? I would separate the TM of wireless to those that are wired. No more than 4 or 5 wireless clients to a target device. Otherwise you are saturating your network with backup data. And use 5ghz.. so close by.. it is much faster than 2.4ghz and you will get much better transfer rates.

How do I install two instances on the same Server and avoid a conflict?

Hi,
I have one Server and I want to install two instances of SQL Server on it.
1. Will be a default instance.
2. Will be a named instance.
The Server has an IP address of 192.168.0.10 and the default instance will be using port 1433.
I want them to share the same address, but be listening on different ports.
Do I,
1. Go into TCP/IP of the named instance and set the port numbers for all the IP's to a port number, e.g. 1976?
2. Delete the 0 from the dynamic ports sections?
3. Create an open port rule in the Windows Firewall to allow access to the Server?
Thanks

Hi,
Please see following Microsoft link
Configure SQL Server to listen on Specific port
Configure Windows firewall to allow SQL Server access
Please mark this reply as answer if it solved your issue or vote as helpful if it helped so that other forum members can benefit from it.
My TechNet Wiki Articles

How can I correlate an Java instance number with a server id?

I have 6 Web App Server running ABAP and the JAVA stacks. I can log into the Central Instance server and run /usr/sap/<SID>DVEBMSG00/j2ee/configtool/configtool.sh and bring up the GUI interface to "see" all the instances and their variables. I don't see an easy way to trace back the Java INSTANCE to the physical server's hostname.
Is there a central repository to correlate a Web App Server Java instance number to the actual hostid/hostname of the target server? I realize I can log into each Web App Server and look at the /usr/sap/<SID>D00/j2ee/cluster/instance.properties and find the Instance number, but I was wondering if this information was held in the SAP Central instance or SAP Database somewhere.

Hi Timothy,
The instance ID used for the JAVA cluster is generated at installation time, and -amongst other things- takes into account the time it was generated. We installed two instances in parallel (on different hosts) and ended up with almost the same instance ID.
Because this can be very confusing, we asked SAP if this was changable - in theory it is only an ID which could be changed as long as it is updated on all places the identifier is stored - but SAP indicated that that is not possible, nor supported. We ended up with a diagram of the system architecture with all the identifiers on our desk.
Note that in the config tool - all instances are displayed with their system instance number. If you were to install a new system it might be usefull to install all instances with different instance numbers so you and distinguish more easily between them.
Hope this helps,
Cheers,
Walter

How to clen up the R12 Instance in Windows 2003 server.

Hi,
We have two R12 instances running on the same server (vision and Development).
and want to scrap Vision instance.
Please explain me how to uninstall the R12 Application instance.
The Note 107523.1: How to Clean Up a Failed Install of Oracle Applications 11.5 on an NT Platform.
But it deletes the entire instances running on the same server.
Thanks and Regards
Vasu

Hi,
Please remove the 3 non-exists DC from environment using the metadata clean-up process using script from below article. then manually delete the records from DNS management console as given below.
https://gallery.technet.microsoft.com/scriptcenter/d31f091f-2642-4ede-9f97-0e1cc4d577f3
Manual steps to clear the records.
Dnsmgmt.msc [Dns Management]
A.Expand the forward lookup zones\_msdcs folder
i. Make sure only the actual domain controllers are listed, delete wrong Alias recordsremove wrong name server records
ii. Select the container [forward lookup zones\_msdcs.domain.com\dc\_sites_\sitename\_tcp] > delete incorrect _ldap and _kerberos records are listed.
iii. Select the container [forward lookup zones\_msdcs.domain.com\dc\_tcp] and delete incorrect _ldap and _kerberos records
iv. Expand the [forward lookup zones\_msdcs.domain.com\domains\guid\_tcp] and delete incorrect _ldap entries
v. Select [forward lookup zones\_msdcs.domain.com\gc] – delete incorrect HostA records
vi. Expand the [forward lookup zones\_msdcs.domain.com\gc\_sites\sitename\_tcp] – delete incorrect _ldap entries
vii.Select the [forward lookup zones\_msdcs.domain.com\gc\_tcp] – delete incorrect _ldap entries
viii. Select the [forward lookup zones\_msdcs.domain.com\pdc\_tcp] – delete incorrect _ldap entries
B.Expand the forward lookup zones\domain.com folder
i.Delete Host(A) records of dc’s which are non-existant.
ii.Correct the NameServer (NS) records
iii. Follow steps similar to ’ A ii ‘ >> ‘ A viii’
· Dssite.msc [Sites and Services]
A.Expand the [Sites\Sitename\Servers] – delete incorrect server’s
B.Delete incorrect subnet configurations [Sites\Subnets]
C.Delete incorrect site links [Sites\IP]
· Make sure the domain controllers are pointing to the correct dns servers in tcp\ip settings.
· Force replication – ‘repadmin /syncall’

How to open two sockets in one single server

May I open two sockets with different listening port numbers in a single server? For each socket, the way to deal with the inputstream and outputstream is different. May the thread be like this:
public class twoSocketServer implements Runnable {
public twoSocketServer() {
try{
SSLServerSocketFactory fact = SSLServerSocketFactory.getInstance(rand, selfPrivateKey, selfCertificate, trustEngine, null);
serverSocket1 = (SSLServerSocket) fact.createServerSocket(ListeningPort1);
serverSocket1.setNeedClientAuth(false);
serverSocket2 = (SSLServerSocket) fact.createServerSocket(ListeningPort2);
serverSocket2.setNeedClientAuth(false);
} catch {...}
public void run()
while(true)
try
SSLSocket socket1 = (SSLSocket) serverSocket1.accept();
new handler1(socket1).start();
SSLSocket socket2 = (SSLSocket) serverSocket2.accept();
new handler2(socket2).start();
} catch {...}
class handler1 extends Thread {...}
class handler2 extends Thread {...}
Is there something wrong with this idea? If yes, how can I correct it? Thanks.

try
SSLSocket socket1 = (SSLSocket)
serverSocket1.accept();
new handler1(socket1).start();
SSLSocket socket2 = (SSLSocket)
serverSocket2.accept();
new handler2(socket2).start();
Is there something wrong with this idea? If yes, how
can I correct it? Thanks.You can do it.
In the above code block you are going to need two threads. One for each SocketServer. The accept() method blocks until it recieves a connection. So in your above code it would first wait only for a connection on port 1. Then it would wait only for a connection on port 2. And then start over. That probably isn't what you want.

How to not run a gateway instance if the target server is down

I have a DirectoryWatcher gateway instance that runs every 10
min looking to do a series of actions when a file is added to a
directory on another server. This other server is down periodically
and my gateway instance fails. Then when the server comes back up,
I have to restart the gateway because it has failed. How do I tell
the gateway instance to not run if the server it is looking at is
down or the directory in this case is unavailable?

listeners have a timeout period. if the target computer is going through a reboot cycle, the downtime should be short enough that you can setup a timeout. Set one that is long enough so that you do not trigger a failure event, and the timepsan allows for the target server to come back online before the timeout expiration.
Another way that is more robust is to create a seperate application, perhaps one built on Visual Studio or even a scripthost that does a simple ping of the destination server every 10 seconds or so. When the server is not found, you can do a shutdown of your CF. (you can shutdown any application from another application). Once you can re-ping the server, relaunch your gateway listener. If the listener cannot support what you want to do, just write another program that will perform your logic for you.
People have it within them to change
Proud user of Adobe products, and webmaster of the Theravive counselor network , built on Dreamweaver, and Photoshop. Toss in a little Visual Studio, but who's keeping score!

Number of 10gR2 instances on a single server

Good evening,
this is my first post here.
We want to create several 10gR2 instances (one per application) on a single linux server (Red Hat 4.0). We do not want to create one instance with separated tablespaces or schemas as the editors of the applications do not support it.
We would like to know how many CPUs we need per instance (in terms of fractional CPU and frequency).
We expect to have from 5 to 15 simultaneous users of these instances (rather OLTP work than decisional). Bigger ones have their own dedicated server.
Do Enterprise edition and Standard edition require the same CPU number/power ?
About multi-core CPUs, is the rule 1dual-core=1.5CPU correct ?
I thank you in advance for your help.

The amount of Oracle Instances you may create in the same box depends, mostly, on the amount of RAM available. You are required to properly configure the unix sempahores, otherwise, even if you have the required free memory, you won't be able to use it.
You have to spread your applications among several instances, but the usage profile is the same among them?, or stated otherwise, are the instances required to be configured just exactly the same? then you can easily estimate the amount of required memory per each instance and the estimated amount of memory required by the users.
Regarding the CPU there is not a thumb rule, you should gather usage statistics based on the current workload present in a typical instance on a typical day (cpu requirements constantly change for the same instance depending on the application usage), and build a baseline after the collected information.
Regarding the kernel parameters (semaphores), I suggest you to continue with this thread where I summarize the semaphore meaning and settings --> Re: Semaphores
~ Madrid

How to install multiple E-Business Suite in a single server?

Hello,
I read somewhere that we can install multiple E-Business Suite in a single server as long as the storage is enough, is it true? If yes do any of you can give me the steps to install it or refer me to a link? I'd searched here but couldn't find it, hope that I didn't miss something. What I want to do is install Vision database and another one is fresh install for testing, both in a single server. There will be about 20 people accessing that system. What kind of hardware must I use? Thanks.

If you want to run 2 instances concurrently, It would be safe to assume you want 8 GB of RAM plus enough storage to host both instances, backups, archive logs, patches etc. For CPU, most entry level servers can be single, dual or quad socket, or CPU socket. Most CPUs are dual core, meaning there are 2 physical CPU cores in each socket. So if you get a 2 socket dual core server, you see 4 physical CPUs on the operating system. If you get a 4 socket dual core server, you see 8 physical CPUs. If you want to run R12 applications, you will definately need 4 GB of ram per instance (E-Business Suite installation). 11i might get by on 3 GB of ram per instance.

How to config RAM of two instance sap on single server ???

Similar Messages

Maybe you are looking for