Database Layout Compromise

I am on a short term assignment at a utility. The database in
question is about 11G in size. It is V 8.1.6 on Open Alpha VMS V
7.2-1. It supports 11 applications, all low OLTP - high query
applications. It is hot backed up/archived. Until yesterday, all
of the files were on 2 (slow) RAID1 volumes(each one disk).
Yesterday, two RAID5 volumes (each volume has 2 disks + 1
parity) were made available to this database. So, I recommended
(for starters) moving (renaming) all of the table data files to
Volume 1 of the RAID5 volumes, index data files Volume 2. This
appears to be working quite nicely. However, all of the other
files (redo, rollback, temp, system) have remained on the old
RAID1 disks. Here is that layout on those two volumes:
Volume3 (Raid 1)
Temporary tablespace #1 datafile
Temporary tablespace #2 datafile
System tablespace datafile
Rollback tablespace #1 datafile
Redo Logs "A" mirror
Volume4 (Raid 1)
Temporary tablespace #3 datafile
Temporary tablespace #4 datafile
Rollback tablespace #2 datafile
Redo Logs "B" mirror
Archived Redo Logs
Please don't ask why there are 4 temp tablespaces! The
explanation I got caused me to lose even more of my hair.
My proposal, given the limited volumes available, is as follows:
RAID5 Volume 1
(This volume currently has all of the table data datafiles.)
1.Move all of the temporary tablespace datafiles over here with
the table data. Obviously, I would not want these with the Index
datafiles. I should mention that all of the temporary
tablespaces are defined as type temporary, but not using
tempfiles. There are very few sorts taking place on disk. That
ratio is excellent. The sort area size is adequate.
2.Move the archived redo logs here as well. (Or would putting
them with the index datafiles be better?)
RAID5 Volume 2
(This currently has all of the index tablespace datafiles.)
No change, unless we put the archived redo logs here, as
mentioned above.
This would leave the mirrored redo logs on Volume 3 and 4
(RAID1), and one rollback segment tablespace datafile on each of
those RAID1 volumes. (Or is is better to put rollbacks on the
RAID5 volumes, and keep Temp tablespaces on the Raid 1 volumes
with the redo logs?) No matter where I put those danged rollback
datafiles, they conflict with something else!)
From my personal experience and reading various books and posts
on datafile layouts, the above recommendation seems to be the
most reasonable compromise to me. I know that it is frowned upon
to put temp or rollback on RAID5, but it is equally frowned upon
to have them on the same RAID1 volumes, especially when redo
logs are also there! I will greatly appreciate all opinions on
possible alternatives!

"List" mode is a specific layout that includes all of the fields, and places them in a spreadsheet-like grid. Your only options are the widths of the columns, the heights of the rows, and the order of the columns.
For a page that displayes only some of the fields, you must create a new layout that contains only those fields. If you want its appearance to be similar to List view, choose a "Columnar Report" as the type of layout to create.
Regards,
Barry

Similar Messages

TIPS(25) : CREATING A DATABASE LAYOUT

제품 : SQL*PLUS
작성날짜 : 1996-11-22
TIPS(25) : CREATING A DATABASE LAYOUT
=====================================
The script below is one I used quite frequently when I was with
Oracle Consulting.
It is handy for walking into a company and getting a quick layout of
their database.
I call it dbmap.sql.
spool dbmap.txt;
prompt
prompt ======================================
prompt Tablespace/Datafile Listing
prompt =====================================
prompt
prompt
column "Location" format A30;
column "Tablespace Name" format A15;
column "Size(M)" format 999,990;
break on "Tablespace Name" skip 1 nodup;
compute sum of "Size(M)" on "Tablespace Name";
SELECT tablespace_name "Tablespace Name",
file_name "Location",
bytes/1048576 "Size(M)"
FROM sys.dba_data_files
order by tablespace_name;
prompt
prompt ======================================
prompt Redo Log Listing
prompt =====================================
prompt
prompt
column "Group" format 999;
column "File Location" format A40;
column "Bytes (K)" format 99,990;
break on "Group" skip 1 nodup;
select a.group# "Group",
b.member "File Location",
a.bytes/1024 "Bytes (K)"
from v$log a,
v$logfile b
where a.group# = b.group#
order by 1,2;
prompt
prompt =====================================
prompt Rollback Listing
prompt =====================================
prompt
prompt
column "Segment Name" format A15;
column "Tablespace" format A15;
Column "Initial (K)" Format 99,990;
Column "Next (K)" Format 99,990;
column "Min Ext." FORMAT 990;
column "Max Ext." FORMAT 990;
column "Status" Format A7;
select segment_name "Segment Name",
tablespace_name "Tablespace",
initial_extent/1024 "Initial (K)",
next_extent/1024 "Next (K)",
min_extents "Min Ext.",
max_extents "Max Ext.",
status "Status"
from sys.dba_rollback_segs
order by tablespace_name,
segment_name;
spool off;

Thanks guys. I'm still trying to learn the programs mentioned. For a person who knows nothing about databases FileMaker was pretty easy to work with. The downside was that it doesn't work with images so well. I was thinking of testing Smart Catalog an extension for InDesign by WoodWing next. Or maybe XCatalog or InCatalog. I'm not too sure Portfolio is right for me, it's great if I wanted to catalog my products with metadata - like a inventory list - but I don't. I want to take the product information from a access/excell file, import it into some type of a catalog program where i can design the layout then save the file as a PDF so I can send it to get professionally printed.
The extension for InDesign might be my best bet, especially since I'm a tad familiar with it. Or Quark, I haven't tried it yet but it seems promising. Thanks for all the suggestions - I'd be lost without all your help.

Proper database layout

I'm looking for some advice on setting up a database with regards to the log files and data files. I come from a SQL Server background and figure some of the basic ideas will follow across to Oracle.
In SQL Server, I would create two partitions on the system, a RAID 1 for my operating system, SQL Server 2005, the master database, and the log files for my user databases. The second would be a RAID 5 partition for the data files for my user databases and nothing more. I assume this is what you'd do in Oracle as well. Is that accurate?
Also, is there a database/schema in Oracle similar to the master database in SQL Server? Is it advisable to move that off to a faster partition in Oracle as it is in SQL Server?
Any information regarding best practices for database layout is greatly appreciated.

the master database in SQL Server is nothing but the SYSTEM tablespace in Oracle which contains all the dictionary and all other database related information. Going with LOCAL MANAGED TABLESPACE and with SEGMENT SPACE MANAGEMENT AUTO are best for automatic management of space. Users tablespace must be created separate from SYSTEM tablespace to relieve contention in database.

Can't get rid of border in database layout

I am creating labels from a database for the Cub Scouts. The database field is surrounded by a thin black border. I can change the thickness of the border, but I cannot get rid of it with the "no line" option. Is it possible?
Appleworks 6.2.9; label has gradient background; font color black.

Dalgor2 wrote:
It turns out that the layout mode puts a border in but that border disappears when in browse mode. I wish they didn't do that as it is confusing, but all is right with the world.
The border around field frames in Layout is to let you see where they are, a very useful feature. The same borders also appear in Browse for individual fields when you place the insertion point in the field. This is also useful, as it initially shows what part of the field contents will be visible when the insertion point is NOT in the field.
Regards,
Barry

Database layout with XML

I'm a java dev, not an oracle admin, so please be gentle!
I've mapped the physical structure of 10 types of devices to 10 XML schemas. These devices have a lot of common attributes (id, location, usage, etc) so I've kept these out of the schemas and, instead, wanted to store them in the DB relationally.
My first instinct was to create a table like:
id varchar2(20),
inventory_xml sys.xmltype
location varchar2(20)
usage varchar2(20)
If I wanted to use Oracle XML DB, though, the "inventory_xml" field does not contain the same type of XML element, because they can be any of 10 different devices, so I can't map the column to a schema. The solutions I came up with are:
1) Wrap the elements in a "device" element so that they are all the same element type. This seems like a hack, however, and goes against the spirit of XML DB.
2) Create a table for each device type, which includes the "id" as a foreign key and the xml data. Then, in the main table, replace the "inventory_xml sys.xmltype" column with "device_type_table varchar2(20)" which stored the name of the correct table.
I have no idea whether this is a good database practice or not, so any suggestions would be appreciated. The reason I want to use XML DB is that the extracts are super fast when you map the xml fields to schemas so queries would take milliseconds instead of 10's of seconds.

garcol wrote:
Hello,
I would like to refresh a region based on a database procedure (it generates html/xml/javascript with SVG).
The layout of the region is generated by the procedure. But when the data changes, the region layout has to change too.
That's why I have made a button with label 'Refresh Region'. I was hoping when a user clicks on it, the region is refreshed. But it does not work (via Dynamic action).
How do I make the region refresh (actually run the procedure again and show the result..).The built-in Refresh dynamic action can only refresh interactive reports, standard reports, charts, and list regions. My initial approach would therefore be to consider rewriting the procedure as a standard report. (I have rarely—if ever—seen a Dynamic PL/SQL region that isn't better implemented as a report with a custom template.)
The other obvious possibility would be to run the procedure as an on-demand process and display/refresh it using an AJAX/JavaScript DA.

How to improve the printing quality of pictures, printing a database layout

I have a database that contains also pictures in multimedia containers. If I print a layout containing a picture, the quality of the picture is really bad! Pixeled with big pixels.
Is there a possibility to get better prints?
The pictures are JPEGs of reasonable size.
The picture in the layout on the screen looks MUCH better.
Printing to an PDF and viewing this on the screen is equally bad.
Any suggestions?
Thanks!

I'm guessing that your images are JPEGs. AppleWorks changes the resolution of all graphic formats except PICT to 72 dpi. There are a couple of things you can do improve the print quality of images.
1. Turn on Fractional Character widths in Preferences > General > Topic: Text. This is a text setting, but it does improve the print quality of images.
2. Open the image in Preview or GraphicConverter & Save As... a PICT to use in AppleWorks.
PICT is generally the best format to use for inserting or pasting into AppleWorks. I find the resulting graphic can be resized somewhat & the resulting saved AW file is much smaller than if using JPEG or TIFF.

Database layout crashes appleworks

good morning. I am using AW 6.2.9 on an iBook G4 w/ OS 10.4.11.
In one of my databases, one layout crashes the program whenever I try to view it in layout mode. I can see in it Browse, and I can see every other layout in both browse and layout. Obviously, it worked at when I first created the layout, so I'm not sure what happened. I've tried creating a new duplicate layout, but it crashes too (but not a new blank one). So other than recreating the whole layout from scratch, can anyone advise why this might be happening and how it can be resolved. Thanks!

Have you added any objects to the layout? The layout editing mode is actually a draw environment & it's possible an added object is corrupt. I don't know how you would be able to delete objects if you can't view it, so I think you might have to recreate it. One thing to try, before going through that, is to save the database as an AppleWorks 5 document & then open this AppleWorks 5 document in AppleWorks 6. There is the slight possibility that this will strip the corruption. You do not need to have AppleWorks 5 to do this.

Database Layout problem

Appleworks 6.0, I'm trying to print a database with only SOME of the existing fields, using specific search criteria. I've saved the searches, created a new layout with only the fields I want to see, but it is only visible in "browse" mode . . . when I try to switch the new layout to "list mode" it shows ALL of the fields!
What am I missing please? THANKS!
Neil
G4 Mac OS X (10.3.9)

"List" mode is a specific layout that includes all of the fields, and places them in a spreadsheet-like grid. Your only options are the widths of the columns, the heights of the rows, and the order of the columns.
For a page that displayes only some of the fields, you must create a new layout that contains only those fields. If you want its appearance to be similar to List view, choose a "Columnar Report" as the type of layout to create.
Regards,
Barry

Need advice on whether to build a traditional database layout or a property bag lookup format for new application?

Hi,
This may be a little of an open ended question, but Id like to ask the crowd as I have a delema to whether or not to use a standard database design or a propertybag/flatfile format and transform the information within my application?
THE SITUATION (EXAMPLE)
We have a table called "people". And these persons can either be a basic contact or be associated to several other "types" such as a member, a representative, a staff member with login details etc. These types are expected to expand
and may need additional fields adding to them.
A: The old fashioned poor approach is to create a table with all the unused columns and simply ignore what we don't need.
B: After normalisation, the traditional approach would be to have link table and each "type" would have its own table with the separate columns needed.
C: Lasty, we have recently done some projects using matrix and property type lookup tables (e.g. a reference table to hold the persons ID, then the columns held in a single properties table of "propertyname and proteryvalue". When the application
"gets" the record it can return a completely dynamic length column list based on how many properties it finds. I.e. Person 1 have only their email address and name in the property bag so the datatable returned in code would be 2-3 columns (inc
id). Person 2 has lots of properties and so their datatable would be returned in code with many more columns.
I hope that makes sense?
THE DILEMA
If I was only using our application to access the data then I could setup a function library to control the access, but I asl want to use report editors (maybe SQL report Services?). And the customer would get completely confused to know which field
is what and how to get the data out? I also want to avoid situations where the application expects someone to have a property (column) and its not there? It may make a difference if I tell you that they will aslo have tables holding heiracrchical
groupings too, which will already make it hard for the customer to use in SQL Report Services. Should I use the traditional approach and the metastable (property bag) approach?
PROS AND CONS of B:
+Traditional approach and most tools that need to access SQL directly will work, assuming the customer understands the column names properly
-A single field change would likely involve all levels of MVC (model, Controller, View) so 3 times the work (maybe 3 people if using specialist SQL Admin, C# Dev and Interface Designer)
-New person types require a whole new table to be linked in
-Global searches need to constantly be updated to include other tables columns
+Our development suite (Visual Studio) allows more visual designing in the IDE when the table is fixed
PROS AND CONS of C:
-Cant see the record easily in SQL directly
-Would have to make any report editor use datasets within an the app rather than direct to SQL directly, making the whole system dependant on a proprietary APP control system (although we would control this anyway)
+Global searches are very good, as all properties are held in a column
-The property table would grow quote big, so needs good indexing
+Any new fields or field groups (types) would simply be a configuration setting in the app (not any manual admin)
+If any presentation of data changes where needed this is likely to only require the controller (developer) on MVC to input changes (assuming the interface expects dynamic fields)
-The tables to show to the user through the app would need to handle dynamic columns and so would need to have generic auto formatting in code (which we could not see in the designer (Visual Studio) and set exactly as required). Slight guess work, but we
could hold setting in the property bag to indicate how to show that property.
A little bit war and peace, but your thoughts?
Kindest
Martin
PS: Some of you may raise the option of stored procedures, in order to do the transform without SQL still. Would this be viable and still work with SQL Report Services?

You ARE using a RDBMS and you appear to intend to use standard development tools. Therefore you will probably regret any attempt to move away from a properly normalized relational database. And that does not begin to address any performance
aspects.
Option A really should not be considered an option. That is not proper normalization for a relational database.
As best I can tell option C is simply an EAV (Entity Attribute Value) model - and you can search on that term for discussions about the pros and cons.
Lastly, AFAIK reports / SSRS can use stored procedures as information sources - that really isn't a design issue. However, if you choose option C then you most likely will be forced to use procedures for reporting purposes.

ERP 2005/UDB 8.2 - Database Layout

I'm looking through the install DVD trying to find the DB layout/install script. Does anyone know when I can find it? (space requirements and layout info).
Thanks!

Hi,
to look at the layout at table level does not make much sense in my opinion. Even from a planning view.
All tables are stored in tablespaces, which are made up by containers on filesystem-level. So the interesting question is: How much size is needed for which tablespace and how to distribute the containers. For a rough space estimation you could look into the appropriate SAP-note (depending on the Release you install, check the installation guide for this note) or look into the file DBSIZE.TPL (R3setup) or DBSIZE.XML (SAPinst), which is located on the first Export Disk of the installation CDs / DVDs.
When you are planning your disk layout be aware not to mix up certain type of files (online-log-files, temp-tablespace, data-tablespaces) on the same disks. Depending on the I/O-rate (SAPS, transactions-rates) of your system you should use enough disk controllers and "spindels".
By the way: Why do you do not use DB2 9?
Kind regards
Waldemar Gaida

Database fields at runtime and freetext search

Hi,
We have started to look into Kodo 3.0 and have used the reversemappingtool
for generating java classes for our current database layout. We are
considering if we should use Kodo/JDO to refactor some of old code.
In our old system we have some features which we would like to move
to the new implementation, but we are not sure if this is possible.
1) At runtime we would like to a database field which we can maintain,
use a default value when inserting and update with new value. Can
it be done, do you have any example code ?
2) In addition to 1, it would be nice if we at runtime could make a custom
table and maintain it.
3) Support for freetext search on Oracle and Sybase, how can it
be done, do you have any example ?
Best regards,
Karsten Mathiasen

Hi Kaushik,
You can make use of APPCC class of your FPM application enhance your appcc class and write your logic to the post exit method. There you will have IO_OVP instance and this interface have all the method which can hide tabs from your application on runtime.
But for the field hide you need to dynamically prepare your screen in your feeder class.
For using APPCC you can refer http://help.sap.com/saphelp_nw74/helpdata/en/3e/741ac912a143368cded7cc418a4bc8/content.htm?frameset=/en/f5/823cdb32254707aec436de005e0ada/frameset.htm.
Thanks
Praveen Gupta

Database Template File missing

Hi everyone,
I'm trying to set up a installation of OCS over multiple computers. The Infrastructre installation works fine withouth any error and all services are up and running. Now I'd like to install and set up the storage database. While running the DBCA, I get an error message stating, that the database layout template file with the extension .dfj is missing. Is this a common problem or could it be that something went wrong while unzipping the file?
The template file for the Infrastructure database is located in both the unzipped installation files AND in the home directory for the Infrastructure.
The installation of the information store also was successfull, but as stated above, i only get an error message while configurating the new database...
Is it possible to tackle the problem in creating a standard new datbase with this assistant?
thank you and happy new year in advance
stefan

Just bumping this thread because I am having the same problem. Is there some configuration setting I missed during installation or can I find this "Transaction_Processing.dfj" file somewhere else?
Thanks for the help.

Appleworks -Viewing Database Records

After creating and typing in several records containing NAMES, ADDRESSES, PHONES, EMAIL, NOTES, I proceeded to print labels with just NAME, ADDRESS (4 Lines). After following instructions on label set up, I then pressed DONE and a message congratulated me on the set up. However, I could not get back to my original pages with the original record with all the information. I was just left with labels. How do I retrieve my original records which was viewabel under "Browse.? Or can I? Was my original data transformed permanently into labels?
Nowhere in the instructions was I told to make a duplicate and saving a copy of my file before making labels. -Frank

Welcome
If you look in the Layout menu, I assumes that at its very bottom you will find at least two items under a separator.
One of these items is the "labels" one.
The other one is the one used to create your records.
Just select it.
It would be useful to search for the keyword "layout" in the help.
Here is an extract:
Creating and working with database layouts
A layout is an arrangement of fields, field names, and other objects. Layouts determine the way data is organized and presented in Browse mode and in reports.
When you create a new database document, AppleWorks automatically creates a layout. You can create other layouts to provide different views of the information in your database. You switch between layouts using the Layout menu or the layout pop-up menu, on the left side of the document window. (The layout pop-up menu does not appear in Layout mode.)
Yvan KOENIG (from FRANCE samedi 5 novembre 2005 20:25:37)

IDES Installation - Database Instance for MS SQL

Hello,
I preveously posted this topic twice in the wrong area of the froum. So I just try again with little bit more knowledge. The problems I had (see history below) seem to be a run time problem (it only took really long). Now it ended up with the error that the startsap.exe could not be started (returncode -1). What could be the reason for that ?
Actually I think I have also a hardware problem. I was not able to find information about the needed HD space. In one SAP note I read that for the IDES on MS SQL 60 GB are needed. I do also have a problem with the topic database file configuration! How this is done in general and what about the files sizes ?
Thanks for any help concerning the installation of the database instance.
BR, Volker
History:
I want to install a IDES 2004 R1 / Netweaver 2004.
I am working on WIN 2003 Enterprise Server with a MS SQL DB. The Central Instance installation ended successfully. I get in trouble with the DB Instance installation:
--> The data to be filled in the sapinst tool was possible, the installation stops at the process step: MS SQL Database Layout. In the log the last Phase is "Setting up Server Phase" and the last two informations after are "Creating Directory", one for data and one for the log. In the explorer they are to see. In the data directory is even a file *.mdf.
Anybody who could tell me what I have to do for the installation ?
Thanks in advance,
Volker
One more note: The information in the footer line of the Installation Wizard says: "Creating database <DB> / this can take a while ...
The while is a never ending while ...
Message was edited by: Volker Lehari8

Hei guys,
the reason why the installation did not continue was the fact that the dispatch+work process didn't started correct. Why it finally did after many reasons I couldn't find out 100% - searching in the SAP notes helped to only try a restart.
So after this I continued (new start of sapinst with "Continue installation") was leading to a success. This was a difficult birth.
Still the question about the needed database space is in my head ! ?? !
BR, Volker

Relational Database & CFM Issues

Issue: I currently have a table in a database that is set up
as such:
1st Column Lists Company Names Following columns lists an a
short hand edition of the type of business each company does in
each state. i.e. In Illnois Comany A provide Type Q Services in
Arkansas Type R etc..
Based on a user selection of a particular state how can I get
the list of CompanyNames and only the selected states information
to display on the page?
CompanyName I IL I AR I AL I CO I
Company A Q R R L
Company B L R Q Q
Company C R R L L
Company D Q L Q Q
Thanks,
P

The database layout has to be updated to do what you want
done.
Rules of Data
Normalization
Table: Companies
CompanyID [ integer ] [primary key] (auto)
CompanyName
Table: BusinessTypes
BusinessTypeID [ integer ] [primary key] (auto)
BusinessType [ varchar ]
Table: CompanyBusinessTypes
CompanyID [ Foreign Key: Companies.CompanyID ]
State [ char(2) ]
BusinessTypeID [ Foreign Key: BusinessTypes.BusinessTypeID ]
This layout allows you to scale up to infinite associations
between companies, states and business types.
SELECT CompanyID, BusinessTypeID FROM CompanyBusinessTypes
WHERE State = '#state#'

Database Layout Compromise

Similar Messages

Maybe you are looking for