QUESTION: Multi-instance memory requirements on Linux

Hi, all.
I've been out of the loop on Oracle technical details for awhile, and need to re-educate myself on a few things. I'm hoping someone can point me to a book in the online docs which discusses my question.
Oracle db 10.2.0.2, on Redhat Linux 2.6.9-67.0.0.0.1. This server is a virtual machine, on a VMWare ESX server.
My question concerns the utilization of memory resources in a multi-instance environment.
I currently have 2 instances/dbs on this server. Each was configured with an SGA_TARGET of approximately 900MB. java_pool_size, large_pool_size and shared_pool_size are also assigned values in the pfile, which I believe supersedes SGA_TARGET.
I am tasked with determining if the server can handle a third instance. It's unclear how much load the database will see, so I don't yet know how much memory I will want to allocate to the shared pool for the buffer cache, etc.
I wanted to see how much memory was being used by the existing instances, so on the server I attempted to capture memory usage information both before, and after, the startup of the second instance.
I used 'top' for this, and found that the server has a total of 3.12GB of physical memory. Currently there's about 100MB free physical memory.
the information from 'top' also indicated that physical memory utilization had actually decreased after I started the second instance:
Before second instance was started:
Mem: 3115208k total, 3012172k used, 103036k free, 46664k buffers
Swap: 2031608k total, 77328k used, 1954280k free, 2391148k cached
After second instance was started:
Mem: 3115208k total, 2989244k used, 125964k free, 47144k buffers
Swap: 2031608k total, 89696k used, 1941912k free, 2320184k cached
Logging into the instance, I ran a 'show SGA', and got an SGA size of about 900MB (as expected). But before I started the instance, there wasn't anywhere near than amount of physical memory available.
The question I need to answer is whether this server can accomodate a third instance. I gather that the actual amount of memory listed in SGA_TARGET won't be allocated until needed, and I also understand that virtual memory will be used if needed.
So rather than just asking for 'the answer', I'm hoping someone can point me to a resource which will help me better understand *NIX memory usage behavior in a multi-instance environment...
Thanks!!
DW

Each was configured with an SGA_TARGET of approximately 900MB. java_pool_size, large_pool_size and shared_pool_size are also assigned values in the pfile, which I believe supersedes SGA_TARGET.
Not quite. If you set non-zero values for those parameters as well as setting SGA_TARGET, then they act as minimum values that have to be maintained before extra free memory is distributed automatically amongst all auto-tuned memory pools. If you've set them as well as SGA_TARGET, you've possibly got a mish-mash of memory settings that aren't doing what you expected. If it was me, I'd stick either to the old settings, or to the new, and try not to mix them (unless your application is very strange and causes the auto-allocate mechanism to apportion memory in ways you know are wrong, in which case setting a floor below which memory allocations cannot go might be useful).
3GB of physical memory is not much these days. The general rule is that your total SGAs should not make up more than about 50% of physical memory, because you probably need most of the other 50% for PGA use. But if your users aren't going to be doing strenuous sorting (for example), then you can shrink the PGA requirement and nudge the SGA allowance up in its place.
At 900MB per SGA, you can have two SGAs and not much user activity. That's 1800MB SGA plus, say, 200MB each PGA = 2200MB, leaving about 800MB for contingencies and Linux itself. That's quite tight and I personally wouldn't try to squeeze another instance of the same size into that, not if you want performance to be meaningful.
Your top figures seem to me to suggest you're paging physical memory out to RAM already, which can't be good!

Similar Messages

  • Coldfusion Multi instance Memory Issues

    Hello,  we recently got brand new servers with 8gb of ram and 64bit Windows 2008.  We have about 7 instances created on these servers and I am noticing something extremely disturbing. On 2 of the instances which I just created today that have absolutely no sites running on them yet as we are still migrating sites.  Coldfusion immediately consumes 700 to 900 mb of working set memory.  this is for all instances which then makes my server seem like it is out of memory.  on the old box it only took the about of working memory that is needed and this would grow over time but not immediately upon starting the server.  I started on of these instances and litterally watched as it took 850mb of ram within 2 minutes of starting and it doesn't get released.
    I do have the jvm set to 2gb and they do for now all share the same jvm.  I am just curious if anyone else running 64bit Windows 2008 is having the same issue and if it is just normal behavior with 64bit systems.  We moved to these beefy servers to help with memory issues but it seems I am still plauged with the issue even when there is no site allocated to the instance.
    Any ideas and thoughts would be appreciated.
    thank you.

    If you have a min memory parameter in your jvm.config (-xmxMin or somethingn like that, then each instance on start will immediately reserve that amount of memory.
    This is a java thing not CF.  With CF8, I don't think having the min memory value matters with the most recent java versions.  I previous versions you would occasionally get out of memory errors if you didn't, but I haven't heard of this since CF8 came out.
    They may have also fixed something in 8 to help alleviate that issue.
    So chaulk it up as normal.  With 7 instances you'll probably run into more CPU issues.
    Byron Mann
    [email protected]
    [email protected]
    Software Architect
    hosting.com | hostmysite.com
    http://www.hostmysite.com/?utm_source=bb

  • Virtualised Multi-Instance SQL Server Cluster - Processor Resource Management

    Hi - We're in the process of implementing a multi-instance SQL 2014 guest cluster on Windows 2012 R2.  To our dismay, it seems that Windows System Resource Manager (WSRM) is deprecated in Windows 2012 R2, so we're now stuck for how best to manage CPU usage
    between SQL instances....
    As far as I can see, I'm left with two options, but both of these have problems:
    1) Use SQL Processor affinity within the guest cluster, with each SQL instance assigned to dedicated v-CPU.  However, I'm not certain that setting SQL Processor affinity within a VM will actually have the desired affect!?..
    - When there is physical CPU capacity available, I'd hope Hyper-V would provide it to whichever v-CPU is demanding it.  
    - When VM processor demand exceeds the physical CPU capacity, I'd hope the SQL instances would receive a proportion of the physical CPU time according to the number of v-CPU(s) assigned through the affinity settings.
    2) Use a VM (actually 2, because its a 2-node guest cluster) per SQL instance!..  This is not ideal, as we need multiple SQL instances and it would result in have an administrative and performance overhead
    Does anyone have any information or thoughts on this?  How can we manage a virtualised multi-instance SQL deployment now that WSRM has been deprecated?  Help me please!

    I'm not sure what are the requirements for each of the 2 VMs in in the SQL guest cluster.
    I'm assuming the guest cluster resides on a Hyper-V CSV with at least 2 Hyper-V hosts, and the 2 VMs are configured with Anti-affinity to ensure they never reside on the same Hyper-V host.
    I've been able to configure CPU resources to VMs from the standard controls in Hyper-V Manager:
    See this blog post
    What edition of SQL 2014 you're using?
    This matters because of these limitations.
    Also consider running SQL Server with Hyper-V Dynamic Memory - see Best Practices
    and Considerations
    Hyper-V performance tuning - CPU
    Hyper-V 2012 Best Practices
    Sam Boutros, Senior Consultant, Software Logic, KOP, PA http://superwidgets.wordpress.com (Please take a moment to Vote as Helpful and/or Mark as Answer, where applicable) _________________________________________________________________________________
    Powershell: Learn it before it's an emergency http://technet.microsoft.com/en-us/scriptcenter/powershell.aspx http://technet.microsoft.com/en-us/scriptcenter/dd793612.aspx

  • About multi instances

    Hi all.
    Very happy I can found this forum.
    My company is considering to use InDesign Server as one of candidate solutions to our Japanese customers.
    I have bassic questions.
    1. Does "multi instances" mean "multi cpu" ? that means, 1 instance for 1 cpu, 2 instances for 2 cpu, 4 instances for 4 cpu ?
    2. If we use multi instances, how InDesign Server will log an error ? each instance will be logged separately to some error files (or another error handling system) ?
    Thank you in advance.

    Multi-instance means running more than one InDesign Server process on the same machine. It is not dependent on the number of processors the machine has. So you could run 3 processes on a single processor machine, etc.
    Each process logs it's errors by writing them to stderr/stdout. You can redirect this whereever you would like. Errors are also logged to an in memory so that they could be accessed from a plug-in.

  • Questions on the Hardware requirement mention in the 12c installation guide

    Hello experts,
    I have a confusion on the Hardware requirements mentioned in the 12c installation guide where they mentioned as below.
    RAM with ADP1, JVMD2 8 GB 12 GB 20 GB
    1 ADP Manager is Application Dependency and Performance
    2 JVMD is JVM Diagnostics
    Question 1: what does this RAM with ADP or JVMD mean. is it just the additional RAM space required in the server or does is that special requirement ??
    Oracle WebLogic Server JVM Heap 1.7 GB 4 GB 8 GB
    Size
    Question 2: What is this JVM Heap size ? how do we configure this in the server ??
    Thanks in Advance for your Help.

    Hi,
    #1: It means if you also want to install ADP and JVMD, the total RAM required is a small (e.g. < 1000 targets), medium (e.g. > 1000 targets and < 10000 targets), large EM deployment (e.g. > 10000 targets) is 8 GB, 12 GB, 20 GB, respectively. That is the memory requirement for an OMS host.
    #2: The heap size for the OMS java process. You can adjust the heap size by changing the -Xmx<heapsize> (see an example below) in the following file:
    <middleware_home>/gc_inst/user_projects/domains/GCDomain/bin/startEMServer.sh
    An example of OMS heap size setting of 2048 MB:
    if [ "${SERVER_NAME}" != "EMGC_ADMINSERVER" ] ; then
    USER_MEM_ARGS="-Xms256m -Xmx1740m -XX:MaxPermSize=768M -XX:-DoEscapeAnalysis -XX:+UseCodeCacheFlushing"
    if [ "${JAVA_VENDOR}" = "Sun" ] ; then
    if [ "${PRODUCTION_MODE}" = "" ] ; then
    USER_MEM_ARGS="-Xms256m -Xmx2048m -XX:MaxPermSize=768M -XX:-DoEscapeAnalysis -XX:+UseCodeCacheFlushing -XX:+UseConcMarkSweepGC -XX:+UseParNewGC -XX:+CMSClassUnloadingEnabled -XX:CompileThreshold=8000 -XX:PermSize=128m"
    Regards,
    - Loc

  • Is there any tool or method to find the exact disk space/memory required for a query to execute in certain time

    Hi,
    We have SSIS Package, which in-turn executes some Stored procedures . 
    This Package runs fine at certain time, say 30 minutes which is acceptable. But sometimes, it takes around 1.5 hr. There was no modification done to the package, and the amount of data handled by the package is also same.
    I have a doubt that some other big process may be running in parallel when this package is getting executed.
    All the Databases are on D drive in the server and currently there is 140 GB free space out of 1 TB.
    Now, using sp_who2 I can check what are the big processes which are running in parallel when this package is getting executed and get their Disk I/O.   
    Now having these information, is there any way to determine, amount of space/memory required to execute a particular query in 30 minutes.
    Or to put a question in other way, I have a query and I know its acceptable execution time(30 min). Now is there any way to determine the criteria to get that execution time.
    NOTE: The query is optimized and required indexes are created. It was running fine, until 1 week back. Suddenly the execution time has increased.
    Can some one please help me in this. Please let me know if you need any further information on this.
    Also whenever these jobs are running, sometimes job activity monitor and SSMS gets frozen. Any idea what does this indicate?
    Thanks in advance,
    Raksha

    Perhaps blocking?
    Take a look at Adam's great stored procedure to see who is running on the server
    Who Is Active? v10.00 (2010-10-21)
    (C) 2007-2010, Adam Machanic
    Feedback: mailto:[email protected]
    Updates: http://sqlblog.com/blogs/adam_machanic/archive/tags/who+is+active/default.aspx
    "Beta" Builds: http://sqlblog.com/files/folders/beta/tags/who+is+active/default.aspx
    License: 
    Who is Active? is free to download and use for personal, educational, and internal 
    corporate purposes, provided that this header is preserved. Redistribution or sale 
    of Who is Active?, in whole or in part, is prohibited without the author's express 
    written consent.
    Best Regards,Uri Dimant SQL Server MVP,
    http://sqlblog.com/blogs/uri_dimant/
    MS SQL optimization: MS SQL Development and Optimization
    MS SQL Consulting:
    Large scale of database and data cleansing
    Remote DBA Services:
    Improves MS SQL Database Performance
    SQL Server Integration Services:
    Business Intelligence

  • Installation of central instance - memory settings

    Hi All,
    I have a question about the memory settings reg. the ABAP / Java Add-In installation (Central system).
    Installation of CI:
    When I install the central instance, I provide a value for the Instance Memory Management.
    Installation of DI:
    In the step where I install the database instance, I provide again a value for Instance Memory.
    Q1: The memory value for the Central instance later is visible in the profile for the CI. But where does the DI value go? How could I change it?
    Q2: The value for the Java JVM (heap size) - is this value additional to the value of the CI, or part of the CI's memory. So could I set heap size to 2M while the CI's memory is set to 1M?
    Any hints are appreciated.
    Thx.
    KB
    System
    - Win2K (Win32)
    - Oracle

    Q1: The memory value for the Central instance later is visible in the profile for the CI. But where does the DI value go? How could I change it?
    --> You can ignore this, there is no INSTANCE memory for your database. If you want to configure the memory settings for your RDBMS software it'll depend on what software it is... For Sql Server it's in the Enterprise Manager in Oracle you can edit the init<SID>.ora file...
    Q2: The value for the Java JVM (heap size) - is this value additional to the value of the CI, or part of the CI's memory. So could I set heap size to 2M while the CI's memory is set to 1M?
    --> This value is specific to your JVM heap, so it's not additive or related to your CI.

  • Cfschedule on multi instance server

    Hi All
    Hope someone can help me on this. We run a multi instance CF server with 3 "slave" instances, which handle the site requests and the CFusion instance that handles the scheduled tasks.
    The application I'm writing has a few scheduled tasks that need setting up each time it gets installed and it is installed multiple times for a variety of clients. So far, this has been done either manually setting them up individually on the CFusiion instance or I created an admin menu extension that fires them off (Knowing the <cfschedule> is being run on CFusion, thus installing them to that instance).
    My question however is; it would much more efficient if the application could check for the existence of the scheduled task and if it's not there, install them. The problem though is as the request would be coming from initial install, thus a web request, it would be coming from one of the "Slave" instances and install the scheduled tasks to them = bad. Is there any way to direct a specific request or point <cfschedule> to install on specific instance of ColdFusion?
    Checking for the existance on the CFusion instance is easy, I can read:
    C:/JRun4/servers/cfusion/cfusion-ear/cfusion-war/WEB-INF/cfusion/lib/neo-cron.xml
    But then directing the request seems to be impossible in this way.
    I am running CF9.0.1 on a Windows 2008 Server box with IIS7.5
    Hopefully somone may be able to help
    Thanks in advance
    Tom

    No one? Bueller? Bueller?
    OK, so I may have come up with a solution, I'm just not sure how elegant it is:
    I have a seperate install.cfm that handles the scheduled task creation. The file that gets included on application start now moves this install.cfm to the /cfide directory and then does a cfhttp call to it on 127.0.0.1 ... which is bound to the cfusion instance ... thus installing the schedule tasks to the correct instance and making my OCD happy. The file is then deleted once the cfhttp returns correct setup.
    Does the job, but as I say, doesnt seem to be the most elegant of ways.
    Tom

  • SQL server client memory required

    Hy
    I want to ask, that I have a MS SQL 2000 Enterprise edition server. How much memory required per client. I know, the SQL system memory required (more 1 G) but my question is, if my SQL server use plus one client, how much memory must take in my blade?
    pls help

    Although there are some thumb rules out there, it is hard to tell as you did not specify which kind of access the users are doing at the database. For ressource intensive querying you should put more memory in your server, where on the other hand you would not have that much memory for users doing plain and simple (Select / DML ) operations. How many users do you have, how many are connected on average, what are they doing on the database ?
    Jens K. Suessmeyer.
    http://www.sqlserver2005.de
    ---

  • OC4J: Determine Memory Requirements for applications

    Hello,
    In the Enterprise Manager Web Site for 9iAS R2, on the System Components page, it displays the CPU% and Memory. Then Clicking on an OC4J instance, it shows the Memory and Heap Usage.
    So, the total memory used by this oc4j instance = memory + heap?
    Is there a documented method to determine memory usage per user for an oc4j application?
    Thanks,
    Manish

    Tuning the JVM depends also on your application.
    In iAS you can add JVM Parameters to the opmn.xml for each OC4J Instance. You have to profile your application in order to determine the correct VM configuration and consider the requirements of the OC4J. Unfortunaly the default configuration is bad. You should add at least add the -sever flag which is available in JDK 1.3.1 used in iAS.
    JDK 1.4.2 offering is offering new features for multiprocessor machines so a good basic stardup would be
    -server -XX:+AgressiveHeap
    which turns on a Parallel Garbage Collector and increase maximum heap size. But there is a point where a single JVM will not be sufficent. Then you have to spwan new JVM processes using the Island concept in iAS.
    I guess this is a good startup guide:
    http://developer.java.sun.com/developer/technicalArticles/Programming/turbo/
    http://java.sun.com/docs/hotspot/gc1.4.2/index.html
    Good look
    Toby

  • InDesign Server CC Multi-Instances

    Following the documentation for multi-instance support the application, running as a web service under IIS7 on Windows Server 2008 R2, is failing to access the object. For a single instance I use CreateObject("InDesignServer.Application") but as document for multiple instances I use GetObject(“myIDS1”) (where I have an instance of InDesign Server CC running with parameter “indesignserver -configuration myIDS1”. I have tried running the instance both from command line, elevated command line, and from the InDesignServerService). It is acting as though it does not recognise the instance name, but then I don't really understand how it would. Any ideas appreciated.

    @anemic.algae, Welcome to the community. If you're looking for a standalone installer, then I would recommend you to download the trial version from the InDesign CC page: https://www.adobe.com/cfusion/tdrc/index.cfm?product=indesign_server. You can later use the prtk tool to serialize the product based on your requirement. Here's the link to how to use the Adobe PRTK tool:
    http://adobe.ly/15VtTWy

  • CUCM 6.X - VmWare Memory Requirement

    Hi there,
    I just installed a CUCM 6.x on a VmWarare and it's seems to works fine except....
    From another PC, I access it by web and it is so slow.. It's pretty much if I can't login because it's too slow.
    What are the physical memory requirement on the PC that I've install the Vm to run CUCM.
    thanks a lot

    Hi Philippe,
    What OS are you running? Have you thought of running Linux?
    Linux would run this more smoothly then windows will especially with so little ram.
    I would set the OS as linux and install VMWare Server. Load the VMWare Image. Back online.
    or
    Upgrade your ram to atleast 1gb. Set your pagefille to 1.5 GB min and 1.5 max with the Windows OS as this will help in stoping your windows os in pagefile allocation.
    HTH, Please Rate
    Regards,
    Justin

  • Multi-Instance BGP

    BGP Multi-instance was supposed to be released in XR 4.2, I could not find any mention of it in the Release notes for 4.2, 4.3, & 5.1?
    How do I find out if it was implemented and if so how to configure it, if not if it is still on the road map?
    Thanks,
    John

    Hello,
    It is in release note XR .4.2:
    http://www.cisco.com/en/US/docs/routers/asr9000/software/asr9k_r4.2/general/release/notes/reln_a9k_42.html
    And some writing on Cisco doc:
    http://www.cisco.com/en/US/docs/routers/crs/software/crs_r4.2/routing/configuration/guide/b_routing_cg42crs_chapter_01.html#concept_E6B1D3D6B8CA4742AED837B3A94E826F
    Below is the example what i did in my lab running 4.2.3:
    RP/0/RSP1/CPU0:ASR-9006-01#show run router bgp
    Fri Jan 31 21:05:02.170 UTC
    router bgp 1 instance IPV4
    bgp router-id 72.15.48.5
    bgp graceful-restart
    ibgp policy out enforce-modifications
    address-family ipv4 unicast
      redistribute connected
      allocate-label all
    address-family vpnv4 unicast
      retain route-target all
    neighbor 172.16.8.3
      remote-as 1
      update-source Loopback0
      address-family ipv4 labeled-unicast
       route-policy PASS in
       route-policy PASS out
    router bgp 1 instance IPV6
    bgp router-id 99.2.3.4
    address-family ipv6 unicast
      redistribute connected
    thanks,
    rivalino

  • Memory Leak in Linux

    Hi friends,
    Is anyone worked on any of the tools to identify and optimise Java Memory Leak in Linux environment?

    Hi friends,
    Is anyone worked on any of the tools to identify and
    optimise Java Memory Leak in Linux environment?Hi,
    I've been using Optimizeit on linux, and I like it.
    /Kaj

  • Application Server Memory Requirements from 4.6C to ECC 6.0

    The projected physical memory increase for the Application servers from 46C to ECC6 appears to between 45% - 66%.  This is based on the delta notes 517085, 752532, 778774 and 901070.  It also appears from a presentation from Christian Oehler called "Transition and Upgrade to SAP ERP" to be the case.
    I cannot find anything in the above referenced SAP notes that indicates where this additional memory is consumed, so I'm concerned SAP is factoring in new functionality and/or JAVA stack in this.  Furthermore we have upgraded our HR system from 46C to ECC6 sometime ago and included JAVA and saw around a 20% memory increase.
    We are doing a strictly technical upgrade and are not turning on new functionality at this point (no JAVA stack), we are also not increasing user population.  Our Test Systems, which have production data volumes do not reflect any growth in this area,  but we are not stressing it with user load.
    Does anyone have an idea of what we could/should see given the nature of our upgrade?  We are preparing for a 25% increase in App server memory, but not for 50% plus.  We are not doing a sizing exercise since this is just a technical upgrade and the GLFU service has yet to start.
    If anyone has any insight into what should be expected or what they have seen for memory requirements from 4.6C to ECC 6.0 Application Server sit would be greatly appreciated.
    Thanks

    Hi Mark,
    More memory is required because ECC 6.0 is much more huge system than 4.6C. If you are planning to increase server memory by 25% only, than you can see performance problems if system will be heavyly used. But it depends mainly on number of WP and active users. 25-50% more memory is required if you are not installing J2EE addin afterwards. If you are planning to convert the database to Unicode, than you will need an additional memory....
    It's not easy to say 25% will be enought or 50% is required, but I would suggest to go to 50% at least.
    Jiri

Maybe you are looking for

  • Can multiple users use Photoshop CC on the same computer?

    My daughter has an account on my MBP and needs to use Photoshop for her Photography class.  This wasn't an issue with Photoshop CS6.  However, with Photoshop CC, it is asking her to purchase a Photoshop CC subscription.  Is there any way she can just

  • How to know if a file browser item is not null using javascript fucntion

    Hi <br><br> I tried to use javascript for validation. I have a file browser item named P55_FILE_NAME and I would like to know if this item is null or not before submit. <br> I wrote this function :<br><br> function validate_import(f_n){<br>      if (

  • Safari - spyglass to open downloaded file in Finder fails

    Just upgraded to 9.0 mavericks - I use Pathfinder 6 - fixed (no more hang). Download in Safari work just fine and end up in the right folder but I can't use the spyglass to highlight a downloaded file in a finder winder I get an error about not findi

  • New Action features

    There are some very nice additions to the functionality of Actions in Keynote '09. One extremely useful improvement that I have yet to see mentioned here is the ability to leave a ghosted image of the final position of an Action on the slide while on

  • Compilation problem with JSP: i am sure this is easy but cant figure out

    this is my jsp file: <@ page import="java.sql.*, com.simpleneasy.*" errorPage="error.jsp"; %> <jsp:useBean id="c" class="ConnectionBean"></jsp:useBean> <html> <%      String sql = "select * from test;";      ResultSet rs;      c.establishConnection()