You want to get in touch with the Oracle Applications Users Group
(OAUG).  Find them at <a
href="http://www.oaug.com/">http://www.oaug.com/</a>.  By the way -
coming up in April is the second annual Collaborate conference, the
co-located conferences of OAUG, the Independant Oracle Users Group
(IOUG), and Quest, the users group for JD Edwards software users.
Last year in Nashville was a fine conference, and this year in Las
Vegas, building on that experience, it should be even better.

--- In VOUGlist@yahoogroups.com, "Sherry" <sobershea@...> wrote:
>
> Hello,
>
> I am looking for an Oracle Financials User Group - where the focus
is
> not the technical [i.e. setup, updates] but rather using Oracle. I
am a
> Financial Analyst with many years of Oracle use but there are still
> things that stump me. We are currently using version 11i.  For
instance,
> I have created many Rules in Oracle [setup, financials, flexfields,
key,
> rules] but am struggling with trying to make one work. It would be
great
> to find a support group!
>
> Does anyone have suggestions?
>
> Thanks in advance,
>
> Sherry
>

As you may've heard, the Ninth Annual SouthEast Oracle User's Conference (SEOUC)
in Charlotte is only six weeks away - this coming March 1st & 2nd, in fact.  The
cool thing about these conferences is that you get two days of solid technical
training for less than what you'd pay for one day at the big West Coast
conferences with all the travel costs.  Plus you don't have to deal with flights
- just get in the car and drive there.  Personally, I'd be happy if I never had
to climb aboard another airplane.

Anyway, there's two things you need to know about this conference.  First,
there's a nifty 10% discount available to all VOUG members - Just choose VIP
passcode from the dropdown registration rate menu and enter the word SPECIAL. 
That brings your final cost down to $220 and some change.  But - you have to do
this *before* February 2nd.

Second, the conference hotel is offering a discounted $135/night rate for SEOUC
attendees who make reservations by January 29th.  This is the Renaissance Suites
Hotel <http://www.renaissancehotels.com/CLTBR> - you won't get that sort of
price from any of the discount travel sites (in fact, I just went to hotels.com
and they quoted me $189/night).  You'll need to call them at 704.357.1414 or
800.HOTELS1 and mention the SEOUC conference to get this discounted rate.

Okay, that's about it.  Well, almost - here's the conference website:
http://www.seouc.com

See you in Charlotte!

Craig

Hello,

I am looking for an Oracle Financials User Group - where the focus is
not the technical [i.e. setup, updates] but rather using Oracle. I am a
Financial Analyst with many years of Oracle use but there are still
things that stump me. We are currently using version 11i.  For instance,
I have created many Rules in Oracle [setup, financials, flexfields, key,
rules] but am struggling with trying to make one work. It would be great
to find a support group!

Does anyone have suggestions?

Thanks in advance,

Sherry

VOUG Membership -  Just a quick reminder - Friday is the the last day
to submit a presentation for the upcoming SEOUC conference.  The
submission date won't be extended, so be sure to submit your
abstract(s) at http://www.seouc.com before it's too late!

As you may've heard, a call for papers is now open for the Ninth Annual
SouthEast Oracle User's Conference (SEOUC) meeting next Spring.  If you've ever
considered presenting at a user group conference - or, if it's been awhile and
you'd like to get back into the user group community - I'd like to encourage you
to give this one a shot.  SEOUC is always an exciting and very well-attended
event.  I always look forward to these conferences.

Now, to those of you who are relative newcomers to Oracle, I would like to
encourage your participation as well.  You see, many times people tend to think
they have to submit some really deep technical paper to get considered.  That is
not the case.  In fact, many attendees would really like to see presentations
that cover the basics. Things like "Here is an overview of what this product
does and here is what we had to go through to get it set up in our shop" or
"Here are the top ten things you need to know to get started with this product."

The point is, if you are working with Oracle products and have tips or advice
you're willing to give, a story to tell, or have figured out some neat way to
"get the job done" then we'd really like you consider sharing it.  One other
thing - this doesn't have to be a one-person show - you are welcome to "tag
team" with other members of your group.  That takes the burden off of you to do
it all alone, plus it's kind of fun.

Complete details about how to submit your abstract can be found at the
conference website:

http://www.seouc.com/

We look forward to receiving your abstract!

Craig

PS.  One more thing - The call for papers is open until next Friday, November
10th.  We won't be extending that deadline, so go ahead and get your submission
in this evening!

OCOJ '06 - Oracle Conference on the James, Williamsburg, VA  November 2, 2006 at
the Great Wolf Lodge.

The final speaker agenda for the Nov 2nd OCOJ conference has been posted on the
OCOJ website - http://ocoj.cc .  Check it out at the 'Schedule' link.  HROUG &
GROUG are very excited with the speaker line up and hope that you will be too.

The conference is only 10 days away!  If any of you out there are waiting for
the last minute to register, please keep in mind that registration closes in 6
days - on Sunday Oct. 29th.  So get your spot reserved today.

The list of goodies that every paid registrant receives more than justifies the
cost of the conference.  In addition to a continental breakfast, box lunch and
goodie filled breaks, each registrant will also receive a book voucher for a
free book of their choice from Rampant Press.  There will be many raffles for
books from O'Rielly and Oracle Press, as well as other vendor goodies to be
given out.

Hope to see you all at the Great Wolf Lodge!

------------------------------------------------------------
Mary Hunsaker Gable
Preferred Training Solutions, Inc
1153 Cordova Court
Virginia Beach, VA  23456
757.721.6655

System
------
2 dual core intel xenon 64 bit
RAM 16g
OS RHEL 4 update 4 64 bit

Database
--------
This database is 99% read only "star schema"
Oracle 10.2 64bit
Our fact table has 17m rows at an avg size of 250bytes
Common queries are

select sum(c1),avg(1),max(1)
sum(c2),avg(c2),max(c2)
from fact
where date between y and z
note this query uses 6 aggregates


Current Settings
----------------
kernel.shmmax=2147483648 ***Seems like this should be 8g but i got an
error when i did that and had to redo my db
kernel.shmmni  = 4096
kernel.shmall  = 2097152
kernel.sem=256 32000 100 142
fs.file-max=131072
net.ipv4.ip_local_port_range=1024 65000
net.core.rmem_default = 1048576
net.core.rmem_max = 1048576
net.core.wmem_default = 262144
net.core.wmem_max = 262144

========================================================

After the above settings are modified, what values would you suggest
for the SGA and PGA?

I feel the biggest slowdown is that the queries being so complex and
spanning random date ranges are filling up the ram and immediately
starting to swap to disk.

My thought process:  We have a view/table whatever we do an aggregate
function on the view Oracle must cache the recordset into memory
containing the rowid and the aggregated columns at a minimum.  If we
figure the fact table has an avg rowsize of 250 bytes the were clauses
and otheroverhead for a few large selects will quickly consume our
sga/pga and we start swapping to disk.  That's where I think the
problem is.

We need to make sure that we have enough memory allocated to Oracle
that we can keep these "temp tables" in memory not swapping.


Steve

Hello,

I would like to start an WVOUG.

Any interest out there?

Thanks,

Jim

Oracle Conference on the James '06 is coming up quickly -- just 5 weeks from
tomorrow -- and early bird registration ends this Saturday, September 30th;
after that the $45 registration fee goes up to $60.

We've had really good response -- seats are filling up quickly (we're at about
50% capacity right now). So if you've been thinking of coming, or waiting for
your manager to approve your request, now's the time to make sure you get
yourself registered at www.ocoj.cc!

Here's some of the speakers and topics you can look forward to enjoying (for
those of you who missed the last GROUG meeting, we're happy to announce that
Barry Johnson has agreed to repeat his excellent presentation at OCOJ as well!)
...

         a.. Robert Bunting -Oracle Corporation
         Oracle Identity Management Suite

         a.. Don Burleson -Burleson Consulting
         Beyond Oracle10g - A look at the future of Oracle in the 21st Century

         a.. Michael Doherty -Network Appliance
         NetApp: Empowering the DBA

         a.. Mary Gable -Preferred Training Solutions
         SQL Developer

         a.. Barry Johnson -World Bank
         Denormalization Debunked in an Hour

         a.. Steve Karam -Burleson Consulting
         tbd

         a.. Tom Kyte -Oracle Corporation
         Oracle 10g: Bits and bytes (2 hours)

         a.. Dave Lapp -Oracle Corporation
         Spatial Capabilities of Oracle 10g (RDBMS & AS)

         a.. Tammi Pierce / Ray Shiniol -US Army
         Semantic Web Using Oracle's RDF Network

         a.. Mark Stock -Enquery
         Change Control -- Just Another Oxymoron?

         a.. Craig Warman -CRT
         Application Development Case Study

Phillip - When Discoverer reports that are included in Oracle Portal,
you can either invoke them as part of the Discoverer portlets (aka
"Drake") interface, or simply link to them as you might from any other
application.  When invoked via the portlet interface, security is
based on the settings you define within those portlets setups, which
essentially means security is being managed by Portal and Discoverer.
  When invoked via link then access to the Discoverer page(s) can be
"protected" through the use of Single Sign-On (SSO).

In either case, Discoverer utilizes database-level security for
determining the workbooks and folders to which a user has access - so
there's a tie-in between the SSO user and the cooresponding database
schema.  Here's a helpful step-by-step tutorial for integrating
Discoverer with Portal:
http://www.oracle.com/technology/obe/obe_bi/discoverer/discoverer_1012/plusporta\
l/Portal.htm

And here's a link to the full documentation set devoted entirely to
publishing Discoverer reports in Portal:
http://download-east.oracle.com/docs/cd/B14099_19/bi.1012/b14465/toc.htm


Concerning the second part of your question, I'd suggest looking at
publishing the Discoverer List of Workbooks page in your portal,
either by providing a direct link to it, or possibly by using the HTML
Clipping portlet.

Hope this helps.

Craig


--- In VOUGlist@yahoogroups.com, "Phillip Small" <phillip.small@...>
wrote:
>
> Hi,
>
> I am working on configuring Oracle Portal for Discoverer reports.
We have a
> J2EE application that also has reports in Discoverer.  We are working on
> implementing Enterprise security to eliminate the users to have to logon
> more than once.  Where is the security managed? Is it in the
> database/Discoverer or the LDAP?
>
> My second question is can I customize the portal so that I can have
folders
> that contain each workbook that they have access to instead of
listing all
> of the individual reports?  I would like to have a main reports page
with
> folder so that the users can click the folder (workbook) which will
bring up
> another page containing the worksheets (reports) in that workbook.
>
> Thanks
>

Hiya All,

I am testing/learning how to make java and oracle work together. My test
system consists of a windows XP based pc that has oracle 10g XE and java
sdk 1.4.2_12. Here is my best version of a first attempt to establish a
connection between java and oracle:



/* Test Oracle 10g XE access with Java */



import java.sql.*;

import java.io.*;



public class queryTest

{ // begin class



   public static void main(String[] args) throws Exception

    { // begin main



     try

     { // begin try



      System.out.println("Loading Oracle driver...");

       Class.forName("oracle.jdbc.driver.OracleDriver");      //
EXECUTION FAILS HERE...



      System.out.println("Initialize connection URL...");

             String url = "jdbc:oracle:thin:@localhost:1521:XE";



      System.out.println("Establishing connection to DB...");

             Connection con = DriverManager.getConnection(url);



      System.out.println("Build SQL Statement object...");

             Statement stmt = con.createStatement();

     } // end try



     catch (Exception e)

     { // begin catch

      System.out.println("Error occurred while building connection to DB:
\n" +

                                                  e.getMessage());

     } // end catch



    } // end main

} // end class









running the code looks like this:



C:\Java>java queryTest

Loading Oracle driver...

Error occurred while building connection to DB:

oracle.jdbc.driver.OracleDriver



C:\Java>





My CLASSPATH environment variable is set to:

.;C:\oraclexe\app\oracle\product\10.2.0\server\jlib\ori18n.jar





I am really hoping that I am doing some dumb thing that will be real
obvious to an experienced db programmer. Please help if you can...

Thanks,

Larry Blankenship

Hi,

I am working on configuring Oracle Portal for Discoverer reports.  We have a
J2EE application that also has reports in Discoverer.  We are working on
implementing Enterprise security to eliminate the users to have to logon
more than once.  Where is the security managed? Is it in the
database/Discoverer or the LDAP?

My second question is can I customize the portal so that I can have folders
that contain each workbook that they have access to instead of listing all
of the individual reports?  I would like to have a main reports page with
folder so that the users can click the folder (workbook) which will bring up
another page containing the worksheets (reports) in that workbook.

Thanks

Hi Everyone,

The HROUG & GROUG user groups are very excited to announce that registration for
OCOJ - the Oracle Conference On The James - is now open.  The conference will be
held in Williamsburg, VA on Thursday, November 2.  We are excited that we have
been able to confirm many high caliber speakers such as Tom Kyte, Don Burleson,
and Mike Abbey, who will be presenting on a variety of technical subjects that
we are sure you will find very educational.  So, make your plans early and come
spend the day networking with other Oracle professionals from around the state
of Virginia.

Visit http://ocoj.cc for conference highlights and registration information.

The Great Wolf Lodge is offering discounted room rentals for those interested in
staying over, and each room rental includes 4 tickets to the indoor water park.

Please pass this announcement on to other Oracle professionals and/or students
whom may or may not be on our mail list. The event is open to every one.  If
your organization is interested in sponsoring, or being a vendor or in
presenting as one of our speakers, please have them contact Hazel Zamperini at
hzamperini@....  We have a very few of each possibly still available.

Has anyone been using Oracle Discoverer?  A group that I work with is
changing from Oracle Browser to Oracle Discover (with our upgradeto Oracle
10g) and are not happy.

1. Is there a tool that is similiar to Browser - maybe a third party -
that you like?
2. Is there anything special about settup in the Discoverer - any little
gotcha's that might be making Discoverer hard to use?

The group is considering that may be they've configured the tool wrong.

Getting another tool is a possibility if, in fact, it's just
Discoverer.....

thanks for any information,
Laureen

P.S.
The group is working with Oracle Support but I thougth you might be able
to help :)

Please verify whether the two tables contain the same data (same column
structure, same rows) or if there are any differences between the two
tables.

Please provide more information on the updates -- how many rows affected,
source of data (OLTP or bulk) # of updates vs inserts, any deletes?
frequency, duration, # of operations. Are the rows being affected grouped
within the same date range or any other logical grouping?

Also, what types of queries -- how likely are they to use rows being updated
(or deleted). How many rows do they include, how long do they run?

The overall size of the tables is not an issue, the transaction load will be
a determining factor. My first reaction is that two tables is very likely
not a good idea, but it depends to a degree on the factors noted above.

-- Mark Stock

----- Original Message -----
From: "Nana Bonsu" <nmbonsu@...>
To: <VOUGlist@yahoogroups.com>
Sent: Friday, August 04, 2006 3:12 PM
Subject: (VOUG) Multi-Table Copies Maintenance


>I just wanted to seek your opinion on a design issue
> that we're trying to resolve.
>
> Basically, users want two tables A and B. Both will be
> about a terabyte in size.
>
> While A is being updated, users will be performing
> queries on B. When the update is complete, users are
> transparently switched from B to A and then B is
> updated to make it as current as A.
>
> Some of the issues we see are the table size and the
> ability to pull this off without impact any
> long-running queries.
>
> We've considered Workspace Manager, Materialized View
> and Synonyms.
>
> Please let me know if you have any ideas about this.
>
> Thanks.
>
> Nana
>

I just wanted to seek your opinion on a design issue
that we're trying to resolve.

Basically, users want two tables A and B. Both will be
about a terabyte in size.

While A is being updated, users will be performing
queries on B. When the update is complete, users are
transparently switched from B to A and then B is
updated to make it as current as A.

Some of the issues we see are the table size and the
ability to pull this off without impact any
long-running queries.

We've considered Workspace Manager, Materialized View
and Synonyms.

Please let me know if you have any ideas about this.

Thanks.

Nana

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

This was a side-effect of the fact that Oracle was doing a GROUP BY
or DISTINCT by sorting the data.  Oracle would have told you even
back then that they don't guarantee that this will always be true.
They would have told you, even then, that the only way to be
absolutely sure of sorting a result set is to use an ORDER BY
clause.  And, frankly, why not use the ORDER BY explicitly?  Seems to
me that the optimiser, even the old rule based optimiser, KNOWS that
it already sorted the data, and it isn't going to do it again.  In
other words, if Oracle sorts for a GROUP BY or DISTINCT, there is
practically no additional performance hit to add the ORDER BY clause.

The only time I ever left off an ORDER BY if I wanted sorted results,
was when I was doing a hierarchical select (START WITH CONNECT BY).
Sorting those results would screw up the hierarchy.  And now, 9i and
10g has ORDER SIBLINGS BY - great stuff.

--- In VOUGlist@yahoogroups.com, "Moorman, Alan" <alan.moorman@...>
wrote:
>
> If you've worked with prior versions of Oracle, at least back to
7.3,
> you're probably used to the output from a SELECT with a GROUP BY or
from
> a SELECT DISTINCT to return the results sorted even when no ORDER
BY was
> specified.  Not so in 10gR2!  Oracle has a note (361549.1) on
MetaLink
> about the GROUP BY issue and I'm still awaiting their response on
the
> SELECT DISTINCT issue.
>
>
>
>
>
> Alan Moorman
>
>
>
> Database Administrator
> Virginia Department of Social Services
> 7 N. 8th Street
> Richmond, VA 23219
>

If you've worked with prior versions of Oracle, at least back to 7.3,
you're probably used to the output from a SELECT with a GROUP BY or from
a SELECT DISTINCT to return the results sorted even when no ORDER BY was
specified.  Not so in 10gR2!  Oracle has a note (361549.1) on MetaLink
about the GROUP BY issue and I'm still awaiting their response on the
SELECT DISTINCT issue.





Alan Moorman



Database Administrator
Virginia Department of Social Services
7 N. 8th Street
Richmond, VA 23219

I was just wondering if anyone used hotcerts for their ocp ?  I am
tempted to try them coz of their price.  Please do let me know.

Hi Everybody,

It is interesting. Oracle adopted Document Object Model (DOM) (worked with
CLOB data type) for XML in Pre 9I version. It introduced XML DB , XMLTYPE
(Data Type) XML extensions for SQL (XMLELEMENT etc)  in 9i onwards. It had
number of problems in 9i related to performance for large files. Oracle
claims more improvement on 10G. In 10G Xquery could handle XML.

In XML DB recomends to use XML schema. It created O-R tables during schema
registration process. We could improve performaces working with O-R
tables/type/collections.

Could you let us know under which framework, you have 220M xml file issue?

How do you load/query XML?

Thanks in advance.

Ram
DW Consultant
http://www.careermag.com/RamSIyer
Ph:410-636-0586


--- In VOUGlist@yahoogroups.com, "wbuker" <william.buker@...> wrote:
>
> Folks,
>
> We are running Oracle Enterprise Edition 64 bit on Solaris Servers.
>
> We have a 220Mg xml file that needs to be parsed and then loaded
> into Oracle tables.
>
> The parse fails by running out of memory. The xml version is up to
> date and the parsing works for files up to 42Mg in size.
> Unfortunatley, the 220Mg file cannot not be broken down into
smaller
> chunks.
>
> Current momory is set to 2.5 Gigs (SETMAXMEMORYSIZE(2500000000)
> ----------------------------)
>
> Any ideas on how to parse this file? Please repond to the email
> address below.
>
> Thanks,
>
> Bill Buker
> Oracle Database Administrator
> Trader Electronic Media
> 757 477-0834 (CELL)
> 757 321-8253 (OFFICE)
> 757-321-8344 (FAX)
> www.TraderOnline.com - Boats, RVs, Collector Cars, Motorcycles and
> More.
> bill.buker@...
>

There are two models for parsing XML, Document Object Model (DOM) and
Simple API for XML (SAX).  DOM parsers have the advantage of letting
you see your XML file as nested objects which lets you quickly access
just the parts of the document you want to see.  However, they need
to read the entire document into memory at once.  SAX parsers act
more like what you might write by hand to parse the document.  They
retrieve the various elements in the document in the order that they
appear in the document.  They only need to have small parts of the
document in memory at any one time.

I suspect that you are using the DOM parser, and you simply don't
have the memory for such a large document to be in memory at once.
If you can switch to the SAX parser, you will probably have better
luck.

--- In VOUGlist@yahoogroups.com, "wbuker" <william.buker@...> wrote:
>
> Folks,
>
> We are running Oracle Enterprise Edition 64 bit on Solaris Servers.
>
> We have a 220Mg xml file that needs to be parsed and then loaded
> into Oracle tables.
>
> The parse fails by running out of memory. The xml version is up to
> date and the parsing works for files up to 42Mg in size.
> Unfortunatley, the 220Mg file cannot not be broken down into
smaller
> chunks.
>
> Current momory is set to 2.5 Gigs (SETMAXMEMORYSIZE(2500000000)
> ----------------------------)
>
> Any ideas on how to parse this file? Please repond to the email
> address below.
>
> Thanks,
>
> Bill Buker
> Oracle Database Administrator
> Trader Electronic Media
> 757 477-0834 (CELL)
> 757 321-8253 (OFFICE)
> 757-321-8344 (FAX)
> www.TraderOnline.com - Boats, RVs, Collector Cars, Motorcycles and
> More.
> bill.buker@...
>

If you are using Oracle Java libraries to parse the
file then check the value for init parameter
JAVA_POOL_SIZE. Probably this value needs be set
higher.

--- wbuker <william.buker@...> wrote:

> Folks,
>
> We are running Oracle Enterprise Edition 64 bit on
> Solaris Servers.
>
> We have a 220Mg xml file that needs to be parsed and
> then loaded
> into Oracle tables.
>
> The parse fails by running out of memory. The xml
> version is up to
> date and the parsing works for files up to 42Mg in
> size.
> Unfortunatley, the 220Mg file cannot not be broken
> down into smaller
> chunks.
>
> Current momory is set to 2.5 Gigs
> (SETMAXMEMORYSIZE(2500000000)
> ----------------------------)
>
> Any ideas on how to parse this file? Please repond
> to the email
> address below.
>
> Thanks,
>
> Bill Buker
> Oracle Database Administrator
> Trader Electronic Media
> 757 477-0834 (CELL)
> 757 321-8253 (OFFICE)
> 757-321-8344 (FAX)
> www.TraderOnline.com - Boats, RVs, Collector Cars,
> Motorcycles and
> More.
> bill.buker@...
>
>
>
>
>
>

Folks,

We are running Oracle Enterprise Edition 64 bit on Solaris Servers.

We have a 220Mg xml file that needs to be parsed and then loaded
into Oracle tables.

The parse fails by running out of memory. The xml version is up to
date and the parsing works for files up to 42Mg in size.
Unfortunatley, the 220Mg file cannot not be broken down into smaller
chunks.

Current momory is set to 2.5 Gigs (SETMAXMEMORYSIZE(2500000000)
----------------------------)

Any ideas on how to parse this file? Please repond to the email
address below.

Thanks,

Bill Buker
Oracle Database Administrator
Trader Electronic Media
757 477-0834 (CELL)
757 321-8253 (OFFICE)
757-321-8344 (FAX)
www.TraderOnline.com - Boats, RVs, Collector Cars, Motorcycles and
More.
bill.buker@...

Hello from a newbie,

I've installed 10g DB on one server running Suse Enterptise and now
I'm attempting to install the Companion CD on another server.  This
server is runing RHEL 3.  Both boxes are nearly identical in hardware
except for a 4 disk RAID in the DB server.
The problem, and it is consistent, is the prerequisite check succeeds
then after selecting either the HTML DB install,first option, or the
HTML DB and Apache install, third option, the runInstaller disappears
and the system sits for hours.  I checked the MD5 of the download of
the Companion CD.  I should point out that I am loading from the HDD
not a CD.

Any hints out there?

   I've gone through this exercise for 3 days now.  I've cleaned out or
removed all directories and also tried leaving them up with no
difference.  Viewing the install scripts indicates that the system
creates the needed directories only if they do not already exists.
When I do not clean the install location, Iget a warning that the
location has content but that does not explain why it fails when there
is not content.

I'm about to abandon Linux and go W2K just to see if I can install
these services that way.  I would do this by swapping out the system HDD.

I'm doing this for a school project and my degree is dependent on this
one project.

I can provide screenshots if you want them.  I just need an eddress to
  send them.

Thanks for any help.
Buddy H

+ Interjection by listserv moderator:
+ Alrighty folks, I think this is what we've all
+ been looking for ;)


Hi all,

   Feel free to revert all the quires about Oracle certification and training to
me.

   regards
   Oracle expert

Hiya All,
I posted a note on VOUG a couple of weeks ago asking for help cleaning
up a linux box after an failed install, so that I could try to
re-install (oracle 10g db on Red Hat Enterprise). I want to thank
everyone who helped me. :)

I am still having troubles however. I have stopped all of the oracle
processes. I have emptied out /home/oracle, deleted the userid oracle,
deleted the /etc/oratab file and deleted the /u01 directory and
everything under it.
Thinking I was able to restart, I went through the pre installation
tasks per the installation guide. Finally I have re-installed, but I get
errors linking the database.

So I ran through the entire process again, and again, I have linking
errors.

Did I forget to delete something?

Could somebody list the steps for "starting over" keeping in mind that I
am both a Linux newbie, and a DBA newbie?

The errors I am getting have the following general format:

Error in invoking target 'racg_install' of makefile
'/u01/app/oracle/oracle/product/10.2.0/db_1/racg/lib/ins_has.mk'
See the log file blah/blah for details

(the target file changes with each error)


Thanks for any help in advance,
Larry

Sjmnetx,

   Hi.

    Ok, when you get a chance to add the line "   set arraysize 1000  ' and re-
run the
query let us know.

   Have you analyzed the table?

   I would think most of the work could be done with indexes but the consistent
gets is making me think otherwise.

Allen


>>> sjmnetx@... 5/1/2006 11:03:32 PM >>>

I had no chance to review the trace yet today.

However,

I checked  dba_indexes and dba_ind_columns for the tables that
comprise all_constraints view;
Namely,  (in SYS. of course)  ('con$' , 'user$' , 'obj$' , 'cdef$') .

The index definitions match on both systems including columns in the
same order.

I'll post more when I get the tkprof the trace.

--- In VOUGlist@yahoogroups.com, "Allen Shepard" <ashepard@...>
wrote:
>
> Sjmnetx,
>   Hi.
>
>   Hmm, well the knee jerk reactions are
>    1) No index to help
>    2) The data in ordered differently between the two tables
causing
> more page fetches.
>
>    Consistent gets is when Oracle goes back to the same untouched
> buffer to get more data.
>
>    If you add the line "   set arraysize 1000  ' and re- run the
query
> see if that helps. You may need to use 500 or 2000 for the array
sized.
> I do not know what the memory setting are for your Oracle.
>
>    Also the page size could be different between the two
installations.
> Check DB_BLOCK_BUFFERS between the two instances.
>
> Hope this helps. "Soo many solutions - so little time. "
>
> Allen
>
>
> Do note my e-mail address is changing to:
> ashepard@...
>
> >>> sjmnetx@... 4/28/2006 3:13:05 PM >>>
>
> Why would a query of  all_constraints cause  7 million consistent
> gets and take nearly two minutes ?
>
> On a similarly configured instance it only takes about 700 CGs and
> 0.02 seconds.
>
> Servers: Oracle 9.2.0.6 on wintel .
>
>
> CRW32.exe  (crystal reports)  generates this query when linking
> tables as it looks for FK relationships.
>
>
> Thanks !
>
>
> SELECT     p_constraint_table.CONSTRAINT_NAME
>           p_constraint_table.OWNER
>           p_constraint_table.TABLE_NAME
>           f_constraint_table.CONSTRAINT_NAME
>           f_constraint_table.OWNER
>          f_constraint_table.TABLE_NAME
> FROM       ALL_CONSTRAINTS f_constraint_table
>           ALL_CONSTRAINTS p_constraint_table
> WHERE
>
f_constraint_table.R_CONSTRAINT_NAME=p_constraint_table.CONSTRAINT_NA
> ME
> AND        p_constraint_table.CONSTRAINT_TYPE='P'
> AND        f_constraint_table.CONSTRAINT_TYPE='R'
> AND        f_constraint_table.OWNER ='GPW'
> AND       ( f_constraint_table.TABLE_NAME='FCI_AGREEMENT'      or
> f_constraint_table.TABLE_NAME='FCI_CONTRACT'  )
> /
>
>
> 1 row selected.
>
>
> Elapsed: 00:01:56.08
>
>
> Statistics
> ----------------------------------------------------------
>           0  recursive calls
>           0  db block gets
>     7738081  consistent gets
>           0  physical reads
>           0  redo size
>         761  bytes sent via SQL*Net to client
>         498  bytes received via SQL*Net from client
>           2  SQL*Net roundtrips to/from client
>           0  sorts (memory)
>           0  sorts (disk)
>           1  rows processed
>
> Thanks Again for any assistance from Oracle land.
>

-- Interjection by listserv moderator:
-- Please email Victoria directly at nichollsvi@...
-- if you interested in responding to this post (don't
-- reply to whole group, in other words...)

Hey guys,

I see that there are a number of you out there with a wealth of experience
(since you spoke up last Allen S.). Steve Karam of Trader has put on some
great presentations in past, especially regarding SQL coding and tuning. How
about a couple of volunteers to speak to people regarding what you know?

I am sure the group would be appreciative.

Victoria

I had no chance to review the trace yet today.

However,

I checked  dba_indexes and dba_ind_columns for the tables that
comprise all_constraints view;
Namely,  (in SYS. of course)  ('con$' , 'user$' , 'obj$' , 'cdef$') .

The index definitions match on both systems including columns in the
same order.

I'll post more when I get the tkprof the trace.

--- In VOUGlist@yahoogroups.com, "Allen Shepard" <ashepard@...>
wrote:
>
> Sjmnetx,
>   Hi.
>
>   Hmm, well the knee jerk reactions are
>    1) No index to help
>    2) The data in ordered differently between the two tables
causing
> more page fetches.
>
>    Consistent gets is when Oracle goes back to the same untouched
> buffer to get more data.
>
>    If you add the line "   set arraysize 1000  ' and re- run the
query
> see if that helps. You may need to use 500 or 2000 for the array
sized.
> I do not know what the memory setting are for your Oracle.
>
>    Also the page size could be different between the two
installations.
> Check DB_BLOCK_BUFFERS between the two instances.
>
> Hope this helps. "Soo many solutions - so little time. "
>
> Allen
>
>
> Do note my e-mail address is changing to:
> ashepard@...
>
> >>> sjmnetx@... 4/28/2006 3:13:05 PM >>>
>
> Why would a query of  all_constraints cause  7 million consistent
> gets and take nearly two minutes ?
>
> On a similarly configured instance it only takes about 700 CGs and
> 0.02 seconds.
>
> Servers: Oracle 9.2.0.6 on wintel .
>
>
> CRW32.exe  (crystal reports)  generates this query when linking
> tables as it looks for FK relationships.
>
>
> Thanks !
>
>
> SELECT     p_constraint_table.CONSTRAINT_NAME
>           p_constraint_table.OWNER
>           p_constraint_table.TABLE_NAME
>           f_constraint_table.CONSTRAINT_NAME
>           f_constraint_table.OWNER
>          f_constraint_table.TABLE_NAME
> FROM       ALL_CONSTRAINTS f_constraint_table
>           ALL_CONSTRAINTS p_constraint_table
> WHERE
>
f_constraint_table.R_CONSTRAINT_NAME=p_constraint_table.CONSTRAINT_NA
> ME
> AND        p_constraint_table.CONSTRAINT_TYPE='P'
> AND        f_constraint_table.CONSTRAINT_TYPE='R'
> AND        f_constraint_table.OWNER ='GPW'
> AND       ( f_constraint_table.TABLE_NAME='FCI_AGREEMENT'      or
> f_constraint_table.TABLE_NAME='FCI_CONTRACT'  )
> /
>
>
> 1 row selected.
>
>
> Elapsed: 00:01:56.08
>
>
> Statistics
> ----------------------------------------------------------
>           0  recursive calls
>           0  db block gets
>     7738081  consistent gets
>           0  physical reads
>           0  redo size
>         761  bytes sent via SQL*Net to client
>         498  bytes received via SQL*Net from client
>           2  SQL*Net roundtrips to/from client
>           0  sorts (memory)
>           0  sorts (disk)
>           1  rows processed
>
> Thanks Again for any assistance from Oracle land.
>

Sjmnetx,
   Hi.

   Hmm, well the knee jerk reactions are
    1) No index to help
    2) The data in ordered differently between the two tables causing
more page fetches.

    Consistent gets is when Oracle goes back to the same untouched
buffer to get more data.

    If you add the line "   set arraysize 1000  ' and re- run the query
see if that helps. You may need to use 500 or 2000 for the array sized.
I do not know what the memory setting are for your Oracle.

    Also the page size could be different between the two installations.
Check DB_BLOCK_BUFFERS between the two instances.

Hope this helps. "Soo many solutions - so little time. "

Allen


Do note my e-mail address is changing to:
ashepard@...

>>> sjmnetx@... 4/28/2006 3:13:05 PM >>>

Why would a query of  all_constraints cause  7 million consistent
gets and take nearly two minutes ?

On a similarly configured instance it only takes about 700 CGs and
0.02 seconds.

Servers: Oracle 9.2.0.6 on wintel .


CRW32.exe  (crystal reports)  generates this query when linking
tables as it looks for FK relationships.


Thanks !


SELECT     p_constraint_table.CONSTRAINT_NAME
           p_constraint_table.OWNER
           p_constraint_table.TABLE_NAME
           f_constraint_table.CONSTRAINT_NAME
           f_constraint_table.OWNER
          f_constraint_table.TABLE_NAME
FROM       ALL_CONSTRAINTS f_constraint_table
           ALL_CONSTRAINTS p_constraint_table
WHERE
f_constraint_table.R_CONSTRAINT_NAME=p_constraint_table.CONSTRAINT_NA
ME
AND        p_constraint_table.CONSTRAINT_TYPE='P'
AND        f_constraint_table.CONSTRAINT_TYPE='R'
AND        f_constraint_table.OWNER ='GPW'
AND       ( f_constraint_table.TABLE_NAME='FCI_AGREEMENT'      or
f_constraint_table.TABLE_NAME='FCI_CONTRACT'  )
/


1 row selected.


Elapsed: 00:01:56.08


Statistics
----------------------------------------------------------
           0  recursive calls
           0  db block gets
     7738081  consistent gets
           0  physical reads
           0  redo size
         761  bytes sent via SQL*Net to client
         498  bytes received via SQL*Net from client
           2  SQL*Net roundtrips to/from client
           0  sorts (memory)
           0  sorts (disk)
           1  rows processed

Thanks Again for any assistance from Oracle land.