SGA Sizing Questions

Hello,
I have a Windows 64-bit server with 16GB of RAM.
Right now, both my sga_max_size and sga_target are set to 4896M.
I've been reading that if I set the SGA too high, it could cause performance issues, but I would still like to increase this a bit.
My database is around 150GB in size, with many reads and writes to the database through HTMLDB applications, also dealing with BLOBS (documents/images stored in our database that are downloaded and uploaded to the system).
I know I can't be told an exact answer, but does anybody have any suggestions on what a safe SGA size would be? I was going to try setting the max_size to 12G and the target to 8G. Does that sound reasonable, or too high/low?
Thanks,
Nora

Check the v$resource_limit to determine if other limits set have been reached.
select * from v$resource_limit;
SQL> /
RESOURCE_NAME CURRENT_UTILIZATION MAX_UTILIZATION INITIAL_ALLOCATION LIMIT_VALUE
processes 999 1000 1000 1000
sessions 93 149 1105 1105
enqueue_locks 34 70 14134 14134
enqueue_resources 34 65 5904 UNLIMITED
ges_procs 0 0 0 0
ges_ress 0 0 0 UNLIMITED
ges_locks 0 0 0 UNLIMITED
ges_cache_ress 0 0 0 UNLIMITED
ges_reg_msgs 0 0 0 UNLIMITED
ges_big_msgs 0 0 0 UNLIMITED
ges_rsv_msgs 0 0 0 0
gcs_resources 0 0 0 0
gcs_shadows 0 0 0 0
dml_locks 0 55 4860 UNLIMITED
temporary_table_locks 0 3 UNLIMITED UNLIMITED
transactions 0 13 1215 UNLIMITED
branches 0 2 1215 UNLIMITED
cmtcallbk 0 2 1215 UNLIMITED
sort_segment_locks 1 9 UNLIMITED UNLIMITED
max_rollback_segments 13 13 1215 65535
max_shared_servers 0 0 UNLIMITED UNLIMITED
parallel_max_servers 0 17 20 3600
22 rows selected.
Regards,
Jess
Edited by: jjhoup on Sep 9, 2008 1:32 AM

Similar Messages

  • Oracle SGA sizing

    Hi,
    I have an oracle rac 10g with 4 nodes on linux and each node has 12G RAM memory. my current SGA sizing from pfile looks like this, I need suggestions to change values to have better performance...
    racdb2.__db_cache_size=2080374784
    racdb3.__db_cache_size=2214592512
    racdb4.__db_cache_size=5620367360
    racdb1.__db_cache_size=2818572288
    racdb2.__java_pool_size=16777216
    racdb3.__java_pool_size=16777216
    racdb1.__java_pool_size=67108864
    racdb4.__java_pool_size=67108864
    racdb1.__large_pool_size=16777216
    racdb2.__large_pool_size=16777216
    racdb3.__large_pool_size=16777216
    racdb4.__large_pool_size=16777216
    racdb2.__shared_pool_size=5737807872
    racdb3.__shared_pool_size=5603590144
    racdb4.__shared_pool_size=2147483648
    racdb1.__shared_pool_size=4932501504
    racdb1.__streams_pool_size=16777216
    racdb2.__streams_pool_size=0
    racdb3.__streams_pool_size=0
    racdb4.__streams_pool_size=0

    SQL> select * from v$sga_target_advice;
      SGA_SIZE SGA_SIZE_FACTOR ESTD_DB_TIME ESTD_DB_TIME_FACTOR ESTD_PHYSICAL_READS
           268           1         471             1            37727
           134          .5       10732          22.7856            54293
           201            .75         498           1.0573            43363
           536           2         435            .9236            29420
           402            1.5         435            .9236            29420
           469           1.75         435            .9236            29420
           335           1.25         442            .9384            31046
    7 rows selected.

  • Catalog database sizing questions

    Hi There,
    We need to create an RMAN catalog that would be able to support 8 databases 5 of which are around 50gb(40GB db files+10gb arcglogs) and the remaing 3 are about 120gb (100 db files and 20 archlogs).
    Does anyone have any sizing recommendations regarding memory structures (sga, pga, large pool)? disk space? We are using Oracle 11.1.0.7 on windows 2008 server. The server itself is an intel box, dual core E5440 - with 16gb RAM.
    Thanks

    Hi rsar001,
    The size of the catalog schema and database in your setup will not grow to more than 1GB (as long as you maintain a retention policy).
    Just set sga_target to 1G.
    As an example I have around 100 databases with sizes bewteen 10 GB and 2T and run a daily backup on all of them retention policy 1 month.
    The catalog is and stays around 700M for more than 2 years.
    Rgds,
    Tycho

  • BW-SEM hardware sizing question

    Hi all, please help me with sizing process
    Now I'm doing sizing of the big BW-SEM system.
    I'm using quicksizer and document "Sizing ASAP FOR BW ACCELERATOR" (https://websmp102.sap-ag.de/~sapidb/011000358700001951182003)
    There is such phrase: "In a comparison of SAP BW benchmark results with SD benchmark results we have obtained a factor of 2.28" (page 11)
    what type of SAPS I obtain from quicksizer - SD or BW?
    So,  should I result from quicksizer tool multiple by 2,28?
    Are there any official documents for this question?
    Thanks in advance,
    Best regards, Elena

    Hi;
    Currently our current 11.5.10 Production is running on a Two node architecture on HP-Unix PA-RISC with the 128 GB of RAM on the Application Server and 256 GB RAM on the DB Node.
    For R12 the servers will be on HP-Unix Itanium and the sizing has been factored as 128 GB for the Application server and the DB will be on RAC with 128 GB of RAM each on the DB Nodes.
    I want to know is the sizing correct?
    When sizing a RAC Env is it ok to just split the current configuration into two and size the same on two nodes?
    Please share your experiences.We do not know how your machine has process which mean we do not know your activiy on your system. You can check some benchmark
    http://www.oracle.com/us/solutions/benchmark/apps-benchmark/ebs-standard-benchmark-overview-192676.html
    Its also important how mch cpu you have.
    Let me give you example from our system. We have 2 node apps server, 2 node rac and db size is almost 18T and our system is working nonstop. Our db has 30CPU and 120g RAm for each node. For apps cpu count and ram count are not like that.10 CPU and 20G ram.
    We are happy wiht apps side, but we still working on better performance issue for our db side
    Regard
    Helios

  • Linux Huge Pages and SGA Sizing

    Hello All -
    I have a RAC cluster (2 node), with 132GB of total memory on each box.
    I am using AMM and have memory_max_target set at 75GB, and memory_target at 70GB.
    I am testing the implementation of huge pages in my lab. My lab cluster has 32GB of total memory, so I set my memory_max_target at 24GB, and memory_target at 22GB.
    After running my baseline test, I saw the SGA was automatically set at 12.75GB, and PGA at 9.25GB.
    My question is as follows:
    I set huge pages to have 24GB of memory locked, so would I set my SGA target at 12.75 and pga_aggregate_target at 9.25GB? Then set sga_max at 14.75GB?
    Any advice is much appreciated!
    Thanks,
    Mike

    Hi Mike,
    little strange, that you can configure huge pages and still use AMM. Normally these features are mutually exclusive.
    MOS Note 361468.1: HugePages on 64-bit Linux
    The AMM and HugePages are not compatible. One needs to disable AMM on 11g to be able to use HugePages. See <<Note 749851.1>> for further information.
    Regards
    Sebastian

  • Thump rule for SGA sizing depends on TOTAL DB SIZE

    Is there any thump rule for sizing SGA depends on TOTAL database size for OLTP system?

    Please dont do that. There was a thumb rule that if you are using oracle >=9i, you can go upto 80% of your physical RAM. But, I will not advice you so.
    Start your SGA with an optimal value, keep on eye on your performance, take statspack report and if further increase required in the SGA, go ahed.
    Because, setting over size SGA cause problems as well setting below par sga also provide poor performance. It should be sized optimially.
    There are few good notes available in metalink.
    Jaffar

  • Needing to add keep pool to SGA, sizing and checking for room?

    Hi all,
    I'm needing to experiment with pinning a table and index (recommended by COTS product vendor) to see if it helps performance.
    I'm trying to set up a keep pool...and put the objects in it
    I've gone into the database, and found that I will need to set up a keep pool:
    SQL> show parameter keep
    NAME TYPE VALUE
    buffer_pool_keep string
    control_file_record_keep_time integer 7
    db_keep_cache_size big integer 0
    That being said, and I'm having a HUGE senior moment right now...how
    do I go about making sure I have enough room to make a little keep
    pool?
    I've looked at my objects I want to put in there, and one is about
    .675 MB, and the other is about .370 MB. So, roughly a little more
    than 1MB
    Looking at my SGA parameters:
    SQL> show parameter sga
    NAME TYPE VALUE
    lock_sga boolean FALSE
    pre_page_sga boolean FALSE
    sga_max_size big integer 572M
    sga_target big integer 572M
    Now...how do I find out what is being used in SGA, to make sure I have room?
    I've been searching around, and trying to come up with some queries. I
    came up with this one:
    SQL> select name, value / (1024*1024) size_mb from v$sga;
    NAME SIZE_MB
    Fixed Size 1.97846222
    Variable Size 232.002007
    Database Buffers 332
    Redo Buffers 6.01953125
    From this, it appears everything is being used....so, not sure what to
    do from here.
    Suggestions and links greatly appreciated!
    cayenne

    SELECT SIZE_FOR_ESTIMATE, BUFFERS_FOR_ESTIMATE, ESTD_PHYSICAL_READ_FACTOR, ESTD_PHYSICAL_READS
      FROM V$DB_CACHE_ADVICE
        WHERE NAME          = 'KEEP'
         AND BLOCK_SIZE    = (SELECT VALUE FROM V$PARAMETER WHERE NAME = 'db_block_size')
         AND ADVICE_STATUS = 'ON';
    SELECT   ds.BUFFER_POOL,
             Substr(do.object_name,1,9) object_name,
             ds.blocks                  object_blocks,
             Count(* )                  cached_blocks
    FROM     dba_objects do,
             dba_segments ds,
             v$bh v
    WHERE    do.data_object_id = v.objd
             AND do.owner = ds.owner (+)
             AND do.object_name = ds.segment_name (+)
             AND do.object_type = ds.segment_type (+)
             AND ds.BUFFER_POOL IN ('KEEP','RECYCLE')
    GROUP BY ds.BUFFER_POOL,
             do.object_name,
             ds.blocks
    ORDER BY do.object_name,
             ds.BUFFER_POOL; Edited by: sb92075 on Jul 9, 2009 2:48 PM

  • SD - situation/effort analysis, sizing questions.

    Hi,
    do not you know if there is some SAP methodology or template which will enable me to get an overview about size of the AS IS situation from sales and distribution point of view in an company?
    I am looking for ideally list of questions which will allow me to estimate effort that will be needed for sucesfull rollout.
    Example of questions coming to my mind....will existing business processes be only used or new have to deployed? How many sales areas (or e.g. shipping points) needs to be created? Etc.
    Please consider this should be an SAP rollout for more distribution center(s).
    Thank you!

    Get a copy of the SAP ASAP methodolgy and use that  or see if you can get a copy of a blue print that has the questions already included and work your way through it.

  • JTree sizing question...

    Hello:
    I have a JTree for which each cell contains a button. I have written a renderer that renders the button, and I realize that I have to do some special stuff to capture a click on the button. My question is unrelated to all of that.
    The problem is that over time, the labels on my buttons change (and may become longer (wider)), but the tree size does not change. In fact, when I update the button label and the tree is re-rendered the rendering of the button gets "chopped off". I've put the tree in a scroll pane, but this doesn't help - the right side of some of the buttons get cut off to the original tree size. I've tried lots of different variations on setPreferredSize, calling repaint, etc, and am not having any luck. I've put together a demonstration of this behavior in a smallish application that I'm posting here, where I create a 2 node tree with buttons that read "Hi", then I change the button labels to "Goodbye" and re-render. You'll see that the button's are cut off about halfway through the button.
    In case its important - I'm running java version 1.5.0_07 on a 32-bit Linux box.
    Any help would be greatly appreciated. Thanks in advance!
    import javax.swing.*;
    import javax.swing.tree.*;
    import java.awt.*;
    public class JTreeQuestion
      public static void main(String [] args)
        JTreeFrame f = new JTreeFrame();
        f.pack();
        f.setLocation(30, 30);
        //Draws buttons with "Hi" (short string)
        f.setVisible(true);
        ButtonNode.updateString("Goodbye");
        //Draws buttons with longer string, buttons get "cut off"
        f.repaint();
    class JTreeFrame extends JFrame
      JTree tree;
      JScrollPane treeView;
      public JTreeFrame()
        super("My Tree");
        DefaultMutableTreeNode root;
        root = new DefaultMutableTreeNode(new ButtonNode());
        root.add(new DefaultMutableTreeNode(new ButtonNode()));
        tree = new JTree(root);
        tree.setCellRenderer(new ButtonNodeRenderer());
        treeView = new JScrollPane(tree);
        add(treeView);
    class ButtonNode
      public static String str = "Hi";
      public static void updateString(String inStr)
      { str = inStr; }
      String getStr()
      { return str; }
    class ButtonNodeRenderer extends DefaultTreeCellRenderer
      public Component getTreeCellRendererComponent(JTree tree,
                              Object value, boolean sel, boolean expanded,
                              boolean leaf, int row, boolean hasFocus)
        Box theBox = new Box(BoxLayout.X_AXIS);
        super.getTreeCellRendererComponent(tree, value, sel, expanded,
                                           leaf, row, hasFocus);
        DefaultMutableTreeNode jtreeNode = (DefaultMutableTreeNode)value;
        theBox.add(new JButton(((ButtonNode)jtreeNode.getUserObject()).getStr()));
        return (theBox);
    }

    For those who are interested. The DefaultTreeModel has a method named nodeChanged() that tells the tree model that a specific node has changed, and the model re-interprets the cell causing its sizse to change as necessary, so that the full button is rendered now.
    Basically what I did was instead of calling repain, I call a method that I wrote that loops through all the tree nodes, indicates they have changed, then repaint's, and it all works out. My trees are relatively small, so this is fine, but if others face the same problem, you'll probably want to selectively indicate which nodes have changed so the tree model doesn't have to do more work than necessary.

  • SGA Sizing

    Dear Experts
    is there any formula to size SGA in 32bit and 64bit. with 4GB or 16GB RAM.
    regards
    saima

    user2108660 wrote:
    Dear Experts
    is there any formula to size SGA in 32bit and 64bit. with 4GB or 16GB RAM.
    regards
    saimaThere is no such formula and do not need AFAIK.But there is main matter is supporting memory like 32bit systems can be support maximum 4G(possible).But in your case you mean is tuning SGA.So its depend your system activities.You have to use Memory Advisor through Enterprise Manager or dynamic performance views(like v$SGA_TARGET_ADVICE).

  • SGA sizing recommendation / white paper

    Hello,
    I am interested in finding a white paper / recommendations on how to best size Oracle database memory SGA. I know it depends on many variables but here is the details that I have so far:
    OS - either windows 2003 or HP unix 11i
    # of databases - 5
    # of schemas per database = 1
    size of each schema 40GB
    # of sessions 30-50 at any given time
    one user defined per schema
    most likely will utilize RAC solution with cluster software on either platform. But we could also use standby database technology instead of RAC as a means of failover/switchover issues.

    I agree with Justin, better to create one instance with 5 schemas then the other way around. This way, you can devote all available system resources, to one instance, and let the apps/schemas achieve a balance within that.
    If you do five instances, you will have to constantly re-balance the system resources manually based on requirements at the time.

  • Sizing Question

    I am using the quicksizer tool and had a question that I could not find the answer to in any quicksizer documentation.
    When you fill out the EP quicksizer active users table there are 4 scenario categories:
         NW-EP-ESS (Employee Self Service)
         NW-EP-INT (EP Intranet scenario)
         NW-EP-PCC (EP People Centric Scenario)
         NW-EP-PRT (EP Custom developement scenario)
    Do I fill out all 4 scenarios or just one. And if a user exists in both do I add them to both?
    Appreciate your assistance,
    Tom

    With the help of a collegue I was able to figure it out. These 4 categroies in the quicksizer are just for notation. It really doesn't matter where you put the data. The only thing you need to fill out is concurrent users, think time, java iviews per page, url iviews per page, % KMC and total users logins within an hour.
    Regards,
    Tom

  • Sizing question (beware newbie)

    Hi all,
    Here’s a question from a true newbie, I hope it is not
    to ignorant to answer though.
    It might be the case that this question has been addressed
    somewhere in this forum already but that I don’t recognize
    the problem due to terminology, to which I’m still to new.
    The other problem is that don’t have the site online so
    it’s not possible to show it live.
    I’ve tried to make an index page (html) with an image
    in the middle like a bar. On this bar a few buttons are placed.
    When I preview this page in a browser it looks right at first but
    when I use the ctrl button and scroll-wheel, (normally altering the
    text size) it completely changes the size of the size except for
    that bar image. The buttons appear somewhere else on the page. (not
    a pretty picture)
    Is there a way to prevent the use of ctrl combined with the
    mouse scroll-wheel?
    It was told me that using fonts measured as pixels this
    wouldn’t be possible but that didn’t prevent this to
    occur.
    Thanks for any help.
    Kind regards,
    Lexz.

    http://apptools.com/examples/pagelayout101.php
    Then move on to these -
    http://www.projectseven.com/tutorials/css/qdmacfly/index.htm
    http://www.macromedia.com/devnet/mx/dreamweaver/css.html
    http://www.macromedia.com/devnet/dreamweaver/articles/tableless_layout_dw8.html
    http://www.macromedia.com/devnet/dreamweaver/articles/css_concepts.html
    Murray --- ICQ 71997575
    Adobe Community Expert
    (If you *MUST* email me, don't LAUGH when you do so!)
    ==================
    http://www.projectseven.com/go
    - DW FAQs, Tutorials & Resources
    http://www.dwfaq.com - DW FAQs,
    Tutorials & Resources
    ==================
    "lexdepax" <[email protected]> wrote in
    message
    news:fn53i6$fet$[email protected]..
    > Ok, that's a clear answer Bregent. Thanks for that.
    > Now as a complete newbie the queston that comes to mind
    is:
    >
    > Is that something one can do visually ( I mean in Dream
    Weaver) or is that
    > always managed from the code itself, which seems to me a
    difficult job.
    > It was so temting to be able to visally place my stuf
    anywhere I wanted
    > it.
    >
    > What would you suggest as the right approach for a
    starter?
    >
    > Thanks again.
    >
    > Regards,
    >
    > Lexz
    >

  • Enterprise DQ - hardware and sizing question

    Hi, we are considering implementing Oracle Enterprise Data Quality (all components) and would like to know the hardware (memory, number of CPU's, etc.) sizing parameters/requirements for this software.  Obviously, data volumes are a key factor.  I'm hoping someone in this community can indicate the "hardware" size they have implemented DQ on and if there are any concerns/issues that, if not considered, can have an adverse impact on the performance and operation of this software stack.
    Thank you.
    Brian.

    Hi Brian,
    See my response to your separate mail.
    Regards,
    Mike

  • CCM - Trex sizing question

    Dear all,
    I'm currenlty installing CCM and Trex on a development system. I was wondering whether someone of you had an indication on what the approximate required disk space is for Trex ?
    The Trex will only be used by CCM and CCM (in our dev system) will have minimal (approx. 100-200) catalog items.
    Thanks for your reply,
    Best regards,
    Kristof

    Hi
    <u>TREX is mandatory for CCM2.0. TREX is a required component of the SAP CCM 2.0, even in the scenario where the search functionality would not be used (i.e. in a scenario of just navigating the hierarchial structure and selecting items) the TREX is still a requirement for publishing and indexing of the published data.</u>
    <b>Other related links of Interest -></b>
    Re: Implement Trex and CCM
    Re: Catalog search (TREX)
    TREX functionnalities
    Re: Configuring trex in ccm 2.0
    Who are the users in SRM sizing?
    Re: sizing the hardware and environment for the upgrade
    Catalog performance issue
    Re: Catalog Mgmt
    Filtering Catalog Search Results
    Configuring trex in ccm 2.0
    Re: Catalog search (TREX)
    Re: Error in Accessing Catalog
    Re: SRM and CCM config
    Re: Are there dependences between the OCI-releases and SRM-releases?
    Re: RFC connections between Iseries (SRM) and Windows 2003 TREX
    Hope this will help.
    Regards
    - Atul

Maybe you are looking for