Configuration of database running on the server with 256 cpu

Hi all

We have a rac 11.2.0.2 on Solaris 10 installation.
The servers are 2 and a short configuration is:

SQL > show parameter cpu_count
VALUE OF TYPE NAME
------------------------------------ ----------- ------------------------------
cpu_count integer 256

psrinfo | WC-l
* 256 *.
prtconf | grep 'memory '.
Memory size: * 130784 MB *.
psrinfo - pv
The physical processor has 64 virtual processors (0-63)
UltraSPARC-T2 + (chipid 0, clock MHz 1414)
The physical processor has 64 virtual processors (64-127)
UltraSPARC-T2 + (chipid 1, clock MHz 1414)
The physical processor has 64 virtual processors (128-191)
UltraSPARC-T2 + (chipid 2, clock MHz 1414)
The physical processor has 64 virtual processors (192-255)
UltraSPARC-T2 + (chipid 3, clock MHz 1414)

There are a lot of settings Oracle automatically design based on the number of CPUs.
In my case, this number is too high.

For example, normally db_cache_size = 4 MB * cpu, in this case the result is 1 G of db_cache_size.

The plan is to create several databases on this car, so the question is:

Is there a way to design or to tell oracle a different number of cpu?
and in general...
You have some tips on how it is possible to configure a machine so powerful that runs a large number of instances?


Thanks in advance

Published by: Spugna on May 7, 2012 03:01

Hi Spugna,

Of course, as long as you use Oracle 11 g you can use Instance Caging to limit each database, the CPU usage. Although it is no longer applicable to the single server database or a CCR node I see no reason why not to use with RAC. Look here for more information:
http://docs.Oracle.com/CD/E14072_01/server.112/e10595/dbrm007.htm

You could also consider the consolidation of these different in a large database of databases if it is possible and there is no dependence application?

Kind regards
EVS

Tags: Database

Similar Questions

  • Script to stop all databases running in the particular ORACLE_HOME

    Hello

    I'm looking for a shell script which would stop all databases running in a particular ORACLE_HOME, then start same accurate databases upward after a 'repeat all' was made.
    Other databases in the particular ORACLE_HOME not running and I don't want to start these. I want only to the stopping and starting again that these databases running in the particular ORACLE_HOME. Thank you.

    I want only to the stopping and starting again that these databases running in the particular ORACLE_HOME.

    Assuming that/etc/oratab is correctly configured and you are working on an already set ORACLE_HOME (otherwise you will need to fix it in scripts).

    $ cat shut.sh
    rm -f /tmp/actives.txt
    for SID in `grep $ORACLE_HOME /etc/oratab | awk -F: '{print $1}'`
    do
            ACTIVE=`ps -ef | grep smon | awk -F_ '{print $3}' | grep $SID`
            if [ "$ACTIVE" ]; then
                    echo $ACTIVE
                    echo $ACTIVE >>/tmp/actives.txt
                    export ORACLE_SID=$SID
                    sqlplus -s / as sysdba << EOF
    shutdown immediate
    exit
    EOF
            fi
    done
    $
    $ cat start.sh
    for SID in `cat /tmp/actives.txt`
    do
            echo $SID
            export ORACLE_SID=$SID
            sqlplus -s / as sysdba << EOF
    startup
    exit
    EOF
    done
    $
    
  • I finished the installation of Sharepoint 2010 and he is told to use the username and password you loged on the server with does not work. What should I do to access the sharepoint URL?

    I finished the installation of Sharepoint 2010 and went through the Setup Wizard. It is said to use the username and password you loged on the server with does not work. What should I do to access the sharepoint URL?

    Hello

    Your question of Windows is more complex than what is generally answered in the Microsoft Answers forums. It is better suited for the IT Pro TechNet public. Please post your question in the TechNet Forum. You can follow the link to your question:
    http://social.technet.Microsoft.com/forums/en-us/sharepoint2010general/threads

  • My Windows 7 Professional Office system disk has been corrupted and I took the opportunity to install a SSD for a new system disk and reinstall Windows.  The system is installed and running on the SSD, with my old drive system currently used for storage a

    My Windows 7 Professional Office system disk has been corrupted and I took the opportunity to install a SSD for a new system disk and reinstall Windows.  The system is in place and running on the SSD, with my old drive system now used for storage as it is still readable.  The only product Adobe, that I reinstalled so far is drive XI but I would like to reinstall items 8 and Acrobat 9 without occupying more licenses that both are also installed on my laptop.  Can you tell me the best way to do it?  -Thank you

    If you have two activations allocated already and one was lost on the hard drive that has failed, then you must contact the Adobe Support via chat and ask them to reset your activations.

    To the link below, click on the still need help? the option in the blue box below and choose the option to chat...

    Serial number and activation support (non - CC) chat
    http://helpx.Adobe.com/x-productkb/global/Service1.html ( http://adobe.ly/1aYjbSC )

  • Unable to connect to the server with ttisqlcs, failed to retrieve IP address of the system

    Please help me, I have installed timesten server and client on different machines in the LAN, but with the same user and group - ttadmin:ttgroup

    When I tried to connect to the server with ttisqlcs - connStr "DSN = sampleCS"

    The result gave me errors like below

    Connect "DSN is sampleCS ';.

    S1000: Unable to retrieve the IP address of the system. System error:-2

    The command failed.

    Fact.

    The related part in sys.ttconnect.ini is defined as below

    [192.168.0.206]

    Description = TimesTen Server

    Adresse_reseau = 192.168.0.206

    PORT_TCP = 53397

    And the sys.odbc.ini

    [ODBC data sources]

    sampleCS = TimesTen 11.2.2 Client Driver

    [sampleCS]

    TTC_SERVER = 192.168.0.206

    TTC_SERVER_DSN = nredb_ds

    That wants it mean by 'recovering intellectual '?

    He is trying to resolve IP by hostname even when I already gave him the ip address?

    Either way, I can ping 192.168.0.206 and telnet to port 53397 without problem.

    Please help me, thanks ~

    It does not matter that they are VMs (I use between virtual machines all the time), but "Networking" of VM, of course, must be set up correctly. But if you can ping and telnet between VMS so that shouldn't be a problem.

    On the client computer, you have an entry in/etc/hosts for the client machines and the server? Even on the server? Please check that and add entries if they are not there and see if it helps. If not then we need to find the client using 'strace' to see which system call fails. Please try and after the release (or at least the last part that leads to failure).

    Thank you

    Chris

  • Change the text of the legend on formready (running on the server)

    Hello

    To change the text of the caption of a field of text as follows:

    (formready, executed on the server, javascript)

    TextField1.caption.value.resolveNode("#text").value = "hello";

    It is a dynamic form and we use LC Forms ES 8.2 on a Windows platform.

    The problem is that the legend IS NOT change the text of the caption, if it runs on THE SERVER. If I change it to run it on the client (for example, Acrobat 9, PDF overview of LC Designer) it does not work.

    Thank you very much for your contributions!

    Daniel

    Set your renderAtClient on no. or false (I don't know the specific parameter). If you leave it in auto, then you allow the form to decide (form properties/Deafults). If the form is dynamic, then by default it will render to the client and not the server.

    Paul

  • Help! HTML updates the server with remote html.1on

    Can help you with this? My html and jpeg files are on the remote server OK. But when I save the changes to the files that they not change or update a new file instead is done on the server with the extension. HTML.1 then I see these two files in the browser. Is this a problem with dreamweaver or my FTP server of the ISP? Here are two files. Any suggestion is appreciated.
    Thank you, Jerry.

    HTML link - the first file uploaded to the server which is not changed / removed from the test server

    link of test html.1 - file that appears on the remote server

    Try to download files using another FTP like IE, WS_FTP client
    FileZilla, to see if the problem is the server or DW:

    http://www.Adobe.com/cfusion/knowledgebase/index.cfm?id=tn_14834#otherclient

    --
    David Alcala
    Adobe technical support

  • script which gives the hostname with more CPU resources and memory available

    Hello

    Anyone have a script which gives the hostname with more CPU resources and memory available in the cluster?

    Concerning

    Vickie

    Hello, VicMware-

    You can get the host with the free CPU, or with the most free memory resources, using the following:

    ## get the host with the most free CPU cyclesGet-Cluster myCluster0 | Get-VMHost | Select-Object Name,    @{n="CpuMhzFree"; e={$_.CpuTotalMhz - $_.CpuUsageMhz}} | Sort-Object -Property CpuMhzFree -Descending | Select -First 1
    
    ## get the host with the most free memoryGet-Cluster myCluster0 | Get-VMHost | Select-Object Name,    @{n="MemGBFree"; e={$_.MemoryTotalGB - $_.MemoryUsageGB}} | Sort-Object -Property MemGBFree -Descending | Select -First 1
    

    The first would have produced something like:

    Name          CpuMhzFree
    ----          ----------
    myVMHost0          25384
    

    And the latter didn't would be out:

    Name           MemGBFree
    ----           ---------
    myVMHost4        122.323
    

    .. .or each of these host computers are those with the most free memory/CPU cluster, respectively.  What to do the things for which you are looking?

  • Home profiles to configure the server with vDS

    Hello

    I'm new to Host Profiles and you want to set up a new generation of server to match the existing reference host. The hardware is identical. I did a manual install and it patched, then I tried that apply the host profile after generating the answer file. I just put IPs in the answer, not the MAC addresses file. I got an error "parameter specified is not correct: PortGroupName. Now the network configuration is a little crooked, the standard vSwitch is still there, and none of the links rising vDS have a physical card associated with them.

    Would someone mind helping me through the steps you would use to apply a profile of the host to a newly built server? I do the distributed configuration manually first vswitch or anything else in preparation? Very much appreciated.

    Thank you

    Mike

    Update: well, I found the source of my error - another admin added a vmkernel port a VDS service and left there for me to stumble, the profile of the host had not been updated since. I deleted and the profile correctly applied this time. My only concern is about to add a new host for the 4 distributed vswitches and migrate the vswitch service VDS Management, I have to prepare all this manually? -Thank you, Mike.

    Hello

    With profiles of the host, you should have a host set up with everything you need (the host of reference). After this, you can create a profile from the host, set this profile to the host you want to modify, verify compliance and then apply the profile. It should ask you unique settings such as the ip addresses of the ports of vmk.

  • Relocation of the oracle database to a new server with upgrade oracle software

    Hi guys,.

    Here's the scenario, we have an existing demo Oracle 9.2 database on Solaris 5.9 server. Asked me this transfer to another server with the same OS but with installed Oracle 10g software. The original plan was to mount the exisiting Oracle home of the original server to the new server and copy there all the databases files and set up the database.

    Here are the steps:
    -Issue a statement "controlfile alter trace database backup", download the script of the trace file
    -Copy all database files to the new server
    -Change the settings files
    -Make a startup mount recreate control files, data recovery and startup of the database on the new server.

    I suggested that we to a tablespace of transport but cause limited resources and that the 9.2 database doesn't have this feature, we cannot carry out this process. Also, I can't make an exp/imp of the database as we really have little resources on a disc. I was wondering if the steps I listed above are correct? Or would result in an error?

    799727 wrote:
    Is it possible not to move the files from the mounted disk and treat it as it is and then upgrade the Oracle 9.2 database to Oracle 10.2?

    Yes you can request the 9.2-10.2 online upgrade, see upgrade guide
    http://download.Oracle.com/docs/CD/B19306_01/server.102/b14238/upgrade.htm#BABJADGA
    http://download.Oracle.com/docs/CD/B19306_01/server.102/b14238/preup.htm#i1007814

  • use the enterprise manager console on the server with 2 instances

    Hello

    I had a server with 2 databases (DB1, DB2), I have 2 windows services (begun boath): OracleDBConsoleDB1 & OracleDBConsoleDB2

    My problem is that I have only 1 url to connect to 1 db (in my case always DB1)

    http://servername:1158/em/console

    is it possible to type the SID in the url, so I can use the company for my 2nd db console?

    (the two db is 11G)

    Kind regards

    O

    orapunk.com wrote:
    Hello

    I had a server with 2 databases (DB1, DB2), I have 2 windows services (begun boath): OracleDBConsoleDB1 & OracleDBConsoleDB2

    My problem is that I have only 1 url to connect to 1 db (in my case always DB1)

    http://servername:1158/em/console

    is it possible to type the SID in the url, so I can use the company for my 2nd db console?

    (the two db is 11G)

    Kind regards

    O

    As suggested at, but not directly stated by others, a given instance of dbcontrol will work with a database. However, you can still have multiple databases on one host, and each database has its own dbcontrol. Differentiation will be each console runs on a different port, and this number is port part of the url that you use to reach the console. The url, with port, is reported when the dbcontrol is configured, and in my view, it is also reported to start the console.

  • Trying to understand HT203325: cannot connect to the server with the account of the user of a network

    It is a kind of home network issues.  Basically what I want to do is set up a unified environment with the profiles of school boards.  I have a couple of iMacs, a couple of the MacBook and some iPods, iPads and iPhones.

    I would like to run the server on one of my iMac, but this note from Apple shows that I'm having problems.  I don't have a machine that I can run like a dedicated server, I need all machines to be accessible by all users of the Open Directory.

    More I read the note, less I seem to understand.  Forgive me, I'm not an IT professional.  I put in place quite a while Apple (10.4 & 10.5) servers and we had no problem with network users that connect on the real server.

    It was that home directories have been shared with AFP or NFS.  I thought that I saw that they are now shared with AFP or SMB.  Note trying to tell if the server is editing on the one hand, that no other client can mount and access a share until the server has removed it?  This seems to contradict the network share point.

    Is there anyway to overcome this problem?  I tried google, but I see a lot of messages that point to the note from Apple, but not really one that helps solve the problem.  Could I run the server on the two iMacs, put on an OD and share the home directories of the other?  So I think that it would be the base directory is never hosted on the same computer where the authentication takes place.

    Any clarification would be appreciated.

    Thank you

    Brett

    PS: My Mac is fairly new, so I can update all 10.11.2 & I didn't buy server, so I would get the most recent and for that, too.

    Try it.   But I wouldn't expect that you are trying to do will be reliable, and I would not recommend using a server as an interactive workstation - the two tasks are usually in opposite directions.

    If you want to store shared files, that activate the OS X client, or install and use a NAS box.   There is no server need, if networked storage is the goal.

    Running a server means implementation of DNS services locally and many other tasks.   It was possible to get out for the most part without doing that, but 10.5 started to get flaky without services local DNS and security and network authentication are now commonplace, and this does not at all well without local DNS.

    Servers should also be available at all times, and - as often happens - shared systems may be rather less stable that software is modified, the systems are closed down, or otherwise.

    NFS is available on OS X for several years, and some people have manually configured access remotely via NFS.   It is much more common to see the AFP and increasingly used CIFS/SMB, however.

    Of HT203325: users can be local, as is typical of most of the OS X client configurations, or they can be so-called network users - users with their connection directory located elsewhere and often with access and the same password on multiple systems.   If users are configured as users of the server network, then do not allow users to connect directly to the server.   Who - is the server that serves files and directories of itself to itself - is not reliable.  Not in this technical note, and not by what I saw.

  • How Auto start the Physical Standby Database (when restarting the server)

    Dear All (s)

    I configure oracle (10.2.0) dataguard on the production database, a time server restart because of disconnectivity to power, I want restard dataguard when rebooting the server. I'm using Server 2003 32-bit. Guide me how I can run after orders when the server restarts.

    SQL > connect/as sysdba
    SQL > bootable media.
    SQL > ALTER DATABASE RECOVER MANAGED STANDBY DATABASE CANCEL;


    Thanks in advance

    I want to run these command each time when the server is restarted.

    Yes, I did not understand that... but what's the problem? You can't schedule a task when the server starts?

  • Not able to connect to the server with modem 56 k

    I try to connect my Tecra via modem 56 k Internet.
    Also, I tried with another computer laptop with the same details and that worked.
    But when I try it on my Tecra I get an error that is not able to connect to the server.

    So I think that it is a parameter that is not defined correctly for the modem.
    (If you compare it to another laptop all parameters are the same as windows).

    Thanks in advance,

    Theo

    Hello Theo

    Please remove the Dial-up connections all created and after that, remove the modem from Device Manager. Restart the device and initially ensure that the modem is properly configured (query modem Agic under Diagnostics). I put t know which Tecra you have, but many of them have preinstalled TOSHIBA PC Diagnostic Tool. With this tool, you can see all the equipment including modem.

    After having it set up new Dial-up connection and check the functionality. Be sure to wait for the tone option is disabled.

  • Trying to access a virtual machine using Blast and VADC configuration without a connection to the server

    Unless I misread this article , it should be possible.

    It is true that step ' not is not particularly clear, but I'm assuming they meant put the content of all unzipped in a folder called 'portal' on the folder root of IIS (C:\inetpub\wwwroot\portal\). In any case I tried this and create a virtual directory in IIS called 'portal' with the wwwroot files, all two complain that there is no such thing as MainUI.html (and it's not! Learners, I can't find this file on a connection or Security Server - which leads me to believe it is generated by one of the .json files)

    Anyone could pull this off? Any suggestions as to where I have gone wrong?

    You can do this by locating the file portal.war on the login server and copy it to the virtual desktop running VADC in view. The procedure is on page 18 of the VADC Administrator's Guide.

    We will make this procedure easier by allowing this portal.war file be downloadable via VMware-Horizon-View-HTML-Access-y.y.y-xxxxxx.zip instead of first needing to install a connection to the server, but for now just copy the file portal.war and unzip it.

    Mark

Maybe you are looking for