Impact of running DBMS_STATS.GATHER_DATABASE_STATS online

Hi,
We're fairly new to DBA with an Oracle 9i database and have been having problems with performance. A lot of the querys were using full table scans instead of indexed scans.
We copied the database to our test database and ran the DBMS_STATS.GATHER_DATABASE_STATS() procedure and lo and behold the queries started using the indices again!!!
We now want to run this procedure via the Oracle Enterprise Manager but are unsure as to the effect that running this procedure will have with our users being logged in at the same time. Is it safe to run this procedure with the users logged on? We are not overly fussed if the users suffer reduced performance in the short term as in the long term they will have much improved performance.
Thanks
Tony

Hi all,
Tony has now gone on holiday for a week. We tried running the DBMS_STATS.GATHER_DATABASE_STATS() via Oracle enterprise manager on the live database last night with no users logged on. it failed with the following errors
ORA-04045 - errors during recompliation/revalidation of sys.dbmd_stats
ORA-06508 - Program being called not found
ORA-06512 - error on line 2
The database and application is supported by a software vendor who claim this is an Oracle bug which is being investigated, the package does however function perfectly on the test database on the same cluster. This test database was rebuilt two days ago using export and import from the live db. I could be wrong but based on some of the claims they have made in the past it seems unlikely. Could this be more of a permissions problem or that the target stats tables are not in place?
What would be the impact of running
ANALYZE TABLE schema.table COMPUTE STATISTICS;
against all tables with users logged on and would this be as effective as the dbms_stats package.
Thanks in advance
Dave.

Similar Messages

  • Error when running dbms_stats.gather_database_stats from dbms_job?

    I had an issue with the default oracle job to analyze the database. I opened a metalink ticket and they said there is a known bug in 10.1.0.3 with the default oracle job and it does not work.
    I am not allowed allowed to update. so the work around was to run this from dba_jobs.
    I have this job created as system and run it manually...
    dbms_stats.gather_database_stats(cascade => TRUE,method_opt => 'FOR ALL COLUMNS SIZE AUTO',
    degree=>8 )
    I can execute this when logged in as system.
    When I try to execute the job I get:
    ORA-12012: error on auto execute of job 26ORA-20000: Insufficient privileges to analyze an object in DatabaseORA-06512: at "SYS.DBMS_STATS", line 12086ORA-06512: at "SYS.DBMS_STATS", line 12425ORA-06512: at "SYS.DBMS_STATS", line 12567ORA-06512: at "SYS.DBMS_STATS", line 12531ORA-06512: at line 1This is running as system?
    The table I don't have access to is located in someones schema.
    log_user,priv_user, and schema_user are all system...
    Why can I run this manually and not through a job?
    Edited by: Guess2 on Jul 21, 2009 9:15 AM
    Edited by: Guess2 on Jul 21, 2009 9:16 AM

    The table I get an error on is not partitioned.
    I also do not get a exception when I run that procedure manually as system, but when I run it from dba_jobs, I get an insufficient privileges issue.

  • Error while trying to run DBMS_STAT.GATHER_TABLE_STATS

    Hi Gurus,
    I am trying to run DBMS_STAT.GATHER_TABLE_STATS from a procedure from a DBA role granted user I am getting the folowing error:
    ORA-20000: Unable to analyze TABLE "<SCHEMA_NAME>"."<TABLE_NAME>", insufficient privileges or does not exist
    ORA-06512: at "SYS.DBMS_STATS", line 13046
    ORA-06512: at "SYS.DBMS_STATS", line 13076
    ORA-06512: at line 2
    Is I am missing something?
    Please help...
    //saby

    It's a quite common problem...
    Privileges granted through roles doesn't work inside stored procedures. So the DBA role it's not working inside procedure.
    See Metalink Note 168168.1.
    If you want to run dbms_stats.gather_table_stats from inside procedure,
    you need to grant "ANALYZE ANY" directly to this user.
    Thanks
    Tomasz K.

  • Bridge CC on opening gives "Adobe Bridge CC has stopped working" Uninstalled, re-installed same error, will not run. "check online for a solution" does nothing.

    Bridge CC on opening gives "Adobe Bridge CC has stopped working" Uninstalled, re-installed same error, will not run. "check online for a solution" does nothing.
    Bridge worked great for a while then started to fail when starting. Old Bridge CS6 still works fine. HELP ... please.

    You problem is in the Fault Module MSVCR80.dll
    Do a web search for this.  Here is one LINK

  • Impact of running dcmctl resetHostInformation

    Hi all,
    Just wondering, what is the exact impact of running dcmctl resetHostInformation ?
    I have a working infra and a working midtier, however, the midtier is not accessible from EM because it's listed as localhost.localdomain. Oracle support claims it was installed this way but it was not.
    So im wondering, can I run emctl resetHostInformation on the midtier only or do I have to run it on infra too?

    Hi Murugesh,
    As mentioned by Giuseppe the PTO carryover process will always recalculate the latest balance, but it is always recommended to follow a cut-off date till which you allow employees to apply and get all the leaves approved. Post that cut-off date you block the previous year leave application in system and run the carry-over process.
    Thanks,
    Sanjay

  • I need to download Firefox v 29 to run a particular online application. Is this possible?

    I need to download Firefox v 29 on a MacBook Pro to run a particular online application (Pearson Q-Global product). How do I find and download an older Firefox version?

    Just in case you aren't sure, those folders are versions in different languages. If you use US English, just download the file in the folder "en-US/".
    After downloading the file, you might also want to verify its sha1 signature against the corresponding entry in [https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/29.0/SHA1SUMS this ]list.

  • Impact of running F.28 / RFDKIL20

    Dear all,
            What would be the impact of running the transaction F.28 or report RFDKIL20? Would thouse released sales order be blocked again?
    Thanks

    F.5D prior to 1KEK
    Dependent functions
        Readjustment must be carried out prior to creating a business area
        balance sheet.
        Profit Center Accounting also requires the results of readjustments.<a href="http://help.sap.com/erp2005_ehp_02/helpdata/en/eb/1376b443c411d1896f0000e8322d00/content.htm">SAP's relevant section of documentation!</a>
    <a href="http://www.sap-basis-abap.com/fico/sap-fi-month-end-and-year-end-activities.htm">Period end activities in order!</a>

  • Is any impact on running instance?

    Hai,
    curently 12.0.4 instance was running on RHEL ES relese 4 ( nanahant 5)  32bit.
    know i am planning to upgrade 32 bit to 64 bit. is any impact on running instance if i upgarde o/s 32 bit to  64 bit.
    Regards
    DBA

    Make sure you relink all executable files/programs once you do the upgrade and before starting the services.
    Note: 471566.1 - Migrating Oracle E-Business Suite R12 from Linux 32-bit to Linux 64-bit
    https://metalink2.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=471566.1

  • Getting error while running script for online backup

    Hi,
    I am running a script for online backup but ended up with an the below error.
    *ERROR* [Backup Worker Thread] com.day.crx.core.backup.Backup Failed to create temporary directory
    Please help out in resolving this.
    Thanks in Advnace.
    Maheswar

    Hi mahesh,
    If you are using backup feature from crx console, I mean http://localhost:4502/crx/config/backup.jsp  I can say that we had also some problems with this functionalities.
    First off all what you need to check are the permissions, because when you check a source code there is line which creates a File object using a path specified by you to make a backup of repository.
    File targetDir = new File(req.getParameter("targetDir", listDir.getParentFile().getAbsolutePath()));
    You need to have sure that the proper read write access has been granted for this path.
    Another issue is that maybe there was already prepared some hotfix if you are using CQ5.4. Please refer to the following link:
    http://dev.day.com/content/kb/home/Crx/CrxSystemAdministration/CRXOnlineBackup.html
    and also to this one:
    http://dev.day.com/content/docs/en/crx/current/release_notes/overview.html which contains a hotfix number #34797 which was applied to backup.jsp file.
    Regards,
    kasq

  • Impact of Running COGI program for every two HRS...

    Hi,
    As per requirement in my business scenario, we have lot of materials with negative stocks currently.
    Now for physical inventory purpose, customer wants us to remove these negative stocks and want us to run COGI program every 2 hrs.
    As we are going to deactivate negative stocks for all materials, the error due to goods movements will get accumulated heavily in COGI.
    For countering this previously they were monitoring COGI once in a day.
    Now we have decided to schedule COGI in background job for every 2HRS.
    Is there any impact on system from PP prospective if we run COGI for 2hrs daily.
    Your view points will be highly appreciated.
    Rgds,
    Yogesh...

    Hello Yogesh,
    it is not possible to schedule COGI (report CORUAFFW) in background.
    But you may schedule report CORUAFWP (which was designed for background running) with option "failed goods movements" in order to clear your COGI records.
    There is no real impact, however, if there are that much records in COGI, you should schedule it for a time with low business in order to avoid lock errors.
    Regards,
    Andreas

  • How to run: system32 dism/online /cleanup-image /restorehealth

    I can run this on my Window's 8.1 pc but not on my Windows 7 pc.

    Hi,
    correct me if I am wrong, but I think that /restorehealth applies only to:
    Windows 8, Windows 8.1, Windows Server 2012, Windows Server 2012 R2
    Please take a look at this link: http://technet.microsoft.com/en-us/library/hh824869.aspx
    I tried DISM on my PC and this is what I am getting when I run: DISM /online /cleanup-image /?
    PS C:\windows\system32> dism /online /cleanup-image /?
    Deployment Image Servicing and Management tool
    Version: 6.1.7600.16385
    Image Version: 6.1.7600.16385
    /Cleanup-Image /RevertPendingActions
      WARNING! This operation will revert pending actions from previous servicing
      operations. It should be used for recovery operations only. This command
      is not supported against an online image.
        Example:
          DISM.exe /Image:C:\test\offline /Cleanup-Image /RevertPendingActions
    /Cleanup-Image /spsuperseded [/hidesp]
      WARNING! The service pack can't be uninstalled after this operation is
      completed. This operation will remove backup files created during service
      pack installation. Use /hidesp switch to hide the service pack from Installed
      Updates.
        Example:
          DISM.exe /Image:C:\test\offline /Cleanup-Image /spsuperseded /hidesp

  • When Running DBMS_STATS inside a Proecure it Fails

    Hello.
    I need to schedule a job to compute stats.
    I decide to create a Procedure GATHERTABLESTATSGTS2 and schedule a job as described below :
    CREATE OR REPLACE PROCEDURE GATHERTABLESTATSGTS2
    as
    BEGIN
    FOR i IN
    select distinct owner,table_name from dba_tables
    where owner in ('GTS2','GTS2_BW')
    and table_name not in ('EXTRACT_ALLOCATIONS_TEU','EXTRACT_TARIFFS_IT')
    LOOP
    BEGIN
    dbms_stats.gather_table_stats
    ( ownname => i.owner      
    , tabname => i.table_name
                   , degree => 4
         , cascade => TRUE);
    EXCEPTION
    WHEN OTHERS THEN
    NULL;
    END;
    END LOOP;
    END GATHERTABLESTATSGTS2;
    VARIABLE jobno number;
    BEGIN
    DBMS_JOB.SUBMIT(:jobno,'GATHERTABLESTATSGTS2;',SYSDATE+2/24, 'SYSDATE+240/1440');
    COMMIT;
    END;
    The Procedure COMPILES WITHOUR ERRORS.
    The job is well scheduled.
    But when I run manually the procedure GATHERTABLESTATSGTS2 I get
    14:17:59 dodbw01.ctr.globedev.ddc> exec sys.GATHERTABLESTATSGTS2;
    PL/SQL procedure successfully completed.
    Elapsed: 00:00:00.93 >>>>>>>>>> executed in less the 1 Seconde !!!!!!!!!!!
    14:18:06 dodbw01.ctr.globedev.ddc>
    And when I check in DBA_TAB_COL_STATISTICS
    select * from DBA_TAB_COL_STATISTICS
    where owner in ( 'GTS2','GTS2_BW')
    order by owner,table_name;
    I don't have any stats , I mean statsitics have not been updated.
    Does someone alredy face this issue before?

    Did you not try what I suggested in your Re: DBMS_STATS Gather_Table_Stats Fails when Called inside a Procedure in the SQL and PL/SQL forum?

  • Poor performance when running DBMS_STATS package

    Hello,
    I thought I would throw this issue out to the general population as I am stumped.
    I have two identical schemas on two different severs. Database in 9.2.0.6 and for QAS and Production reasons the servers and database settings are identical.
    I am running the DBMS_STATS.GATHER_SCHEMA_STATS package on both databases. On the QAS DB it takes roughly 1.5 hours while on the Production database it takes 8.0 hours.
    Looking at the statspack analysis, the biggest thing to jump out at me is that the temporary tablespace is getting 4.6 Million I/O hits (reads) while the QAS is only getting a few thousand.
    So .... this sort of indicates that something is going on in the PGA which could cause Oracle to use the TEMP tablespace. However when I look in the host performance (using Grid Control) the servers memory, file I/O, and CPU while high are not over 80%.
    Has anyone else seen and issue like this? Any ideas where to look?
    Thanks,
    Chris

    To the extent that you are gathering histograms on the data, Oracle potentially has to do a reasonable amount of sorting. If one machine has substantially more physical RAM, that presumably implies that it has more RAM available for the PGA, which would allow more sorts to happen in memory rather than on disk.
    Depending on how you create a "refresh copy of production" and what parameters you are passing in to DBMS_STATS, it is also possible that the quality assurance database is gathering substantially more histograms than the production database (i.e. you refresh lower environments via export and import rather than doing an RMAN clone and tell Oracle to gather histograms on any columns that currently have histograms).
    Justin

  • Impact of running autoconfig on SSL setup

    Hi,
    I have SSL setup in production and need to apply the latest autoconfig patch there. Can you please advice if running autoconfig will impact the current SSL setup in any way?
    My EBS version is 11.5.10.2 with 10.2.0.4 database. I have a multinode configuration with Web and forms server on one node and Database, reports, concurrent and admin server on the other.

    user503988 wrote:
    Hi,
    I have SSL setup in production and need to apply the latest autoconfig patch there. Can you please advice if running autoconfig will impact the current SSL setup in any way?
    My EBS version is 11.5.10.2 with 10.2.0.4 database. I have a multinode configuration with Web and forms server on one node and Database, reports, concurrent and admin server on the other.AutoConfig should not impact your SSL setup, as long as you followed these MOS docs to implement SSL and have all the context variables set properly.
    11i: A Guide to Understanding and Implementing SSL for Oracle Applications [ID 123718.1]
    11i: Troubleshooting SSL with Oracle Applications [ID 300969.1]
    Thanks,
    Hussein

  • What is the possible impact of running BDLS in wrong client?

    Hello,
    I've been doing a system refresh from PRD to QUA and as a part of post refresh actions I've run BDLS for each client I have in the system. That system has 4 clients (500,600,700,800) so I've run BDLS 4 times (PRDCLNT500 -> QUACLNT500, PRDCLNT600 -> QUACLNT600, PRDCLNT700 -> QUACLNT700, PRDCLNT800 -> QUACLNT800). The problem is that I've run all four reports from the same client (500).
    When BDLS finished and I've been checking the spool reports I found out my mistake - there were over 5.000.000 updates in correct report (PRDCLNT500 -> QUACLNT500) and only 3.000 updates in wrong ones (all other reports).
    After this I've started BDLS for the remaining clients (600,700,800) correctly.
    <b>The question is - could there be any possible impact on data consistency?
    If so - what impact could I expect?
    Would be wise to run the BDLS reports in inverse mode (changing the logical names to the original ones) in client 500?</b>
    Here is some information that can help - tables with amount of modified records:
    BDSX_CON04*                    LOG_SYSTEM                             1                      1
    ROIDOCPRMS*                    SLOGSYS                                1                      1
    ROOSGEN*                       RLOGSYS                                0                      0
                                   SLOGSYS                                1                    201
    ROOSPRMSC*                     RLOGSYS                                0                      0
                                   SLOGSYS                                1                      3
    ROOSPRMSF*                     RLOGSYS                                0                      0
                                   SLOGSYS                                1                      4
    RSBASIDOC*                     RLOGSYS                                0                      0
                                   SLOGSYS                                1                      1
    SRRELROLES                     LOGSYS                                 1                  2.040
    TBTCO*                         RECLOGSYS                              1                      1
    EDIDC                          RCVPRN                                 0                      0
                                   SNDPRN                                 1                    780
    EDP21                          SNDPRN                                 1                      3
    EDPP1                          PARNUM                                 1                      1
    T076B                          KONTO                                  1                      1
    T076M                          KONTO                                  1                      1
    T076S                          KONTO                                  1                      1
    I expect that all tables marked with star (client independent tables) have no impact (they would be changed anyway).

    Thanks. I got similar response from other SAP experts I've consulted. But I do not understand what are the changed records? If there would be no changed records - I would agree there is no impact. But when I've by accident changed 3x3000 records - it is hard to belive that there is absolutely no impact.
    Can you please explain?

Maybe you are looking for

  • IPOD nano not detected by updater software or Itunes

    Well, what a struggle. I've spent the last 3 nights trying to get this inanimate lump of plastic to work, which I bought earlier this week. Synopsis is that windows can see the ipod, but neither Itunes nor the Ipod updater can. Frustratingly I had a

  • How can i feed a text file into a Hashtable?

    Hi, im working on my third year project, im new to java and am struggling a little with writing the code. I need to feed a text file into my program and insert it into a hashtable. I presume i'll be using an input stream to feed the file in, how do i

  • Regarding Smartforms Main Window

    Hi SAP Gurus, In a customized smartform, i am having an output  : srl no                 mat no                  mat desc                     quan              uom               rate 10                        123456             aaaaaaaaa             

  • Dynamic view names in view objects (adf faces bc 10g)...

    I have written an application in adf faces and business components, but would like to be able to add sections to it by simply creating a database view, say with standard names and a suffix, and probably static number of columns (just a read only view

  • Export/Import error : ORA-31609: error loading file "kualter.xsl"

    Hi , I was exporting the database , I got the following error after that I found that dbms_metadata_util.load_stylesheets needs to run . Please assit me how to procced further . C:\Users\379164>expdp import1/import1 directory=dmp_dir dumpfile=dump.dm