Best practices for an oracle application upgrade

Hello,
We have an enterprise application deployed on Oracle Weblogic and connecting to an Oracle database (11g).
The archive is versioned and we are using Weblogic's feature to upgrade to new versions and retire old versions.
In a case of emergency when we need to rollback an upgrade, the job is really easy on Weblogic but not the same on Oracle DB.
For most of our releases, the release package is an ear plus some database scripts.
Releases are deployed with minimum downtime, so while we are releasing our clients are still writing to the DB.
In case of a rollback is needed, we need to make sure the changes we made to the DB structure (Views, SP, Tables...) are reverted but data inserted by clients stayed intact.
Correct me if I am wrong, but Flashback and RMAN TSPITR are not the good options here.
What other people usually do in similar cases? What are best practices and deployment plans for our case?
Guides and direction are welcomed.
Thanks!

Hi Magnus
I guess you have to install again to ensure no problems. BP installation also involves ensuring correct SP levels (cannot be higher) for all software components.
Best regards
Ramki

Similar Messages

Best practice for integrating oracle atg with external web service

Hi All
What is the best practice for integrating oracle atg with external web service? Is it using integration repository or calling the web service directly from the java class using a WS client?
With Thanks & Regards
Abhishek

Using Integration Repository might cause performance overhead based on the operation you are doing, I have never used Integration Repository for 3rd Party integration therefore I am not able to make any comment on this.
Calling directly as a Java Client is an easy approach and you can use ATG component framework to support that by making the endpoint, security credentials etc as configurable properties.
Cheers
R
Edited by: Rajeev_R on Apr 29, 2013 3:49 AM

Best practice for install oracle 11g r2 on Windows Server 2008 r2

Dear all,
May I know what is the best practice for install oracle 11g r2 on windows server 2008 r2. Should I create a special account for windows for the oracle database installation? What permission should I grant to the folders where Oracle installed and the database related files located (datafiles, controlfiles, etc.)
Just grant Full for Administrators and System and remove permissions for all others accounts?
Also how should I configure windows firewall to allow client connect to the database.
Thanks for your help.

Hi Christian,
Check this on MOS
*RAC Assurance Support Team: RAC Starter Kit and Best Practices (Windows) [ID 811271.1]*
https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&doctype=BULLETIN&id=811271.1
DOC Modified: 14-DEC-2010
Regards,
Levi Pereira

Best practice for installation oracle 11g rac on windows 2008 server x64

hello!
can somebody tell me a good book or an other kind of literature regarding "best practice for installation oracle 11g rac on windows 2008 server x64"? thx in advance!
best regards,
christian

Hi Christian,
Check this on MOS
*RAC Assurance Support Team: RAC Starter Kit and Best Practices (Windows) [ID 811271.1]*
https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&doctype=BULLETIN&id=811271.1
DOC Modified: 14-DEC-2010
Regards,
Levi Pereira

Best practice for mouseless ADF applications

I am developing an ADF application where the users do not want to use the mouse.
So I would like to know if there are a best practice for this?
I am already using the accessKey functionality and subforms defaultCommand
But I have had problems setting focus to objects on a page like tables. I would like a button to return the focus to the table after it has made the command like delete.
I have implemented a solution where I have found inspiration several threads and other webpages (see below).
Is this solution okay?
Are there any problems with it?
I would also like to know if there are better pathways to go like
out of the box solutions,
http://www.oracle.com/technetwork/developer-tools/adf/learnmore/79-global-template-button-strategy-360139.pdf (are there an example implementation?), or
http://one-size-doesnt-fit-all.blogspot.dk/2010/11/adf-ui-shell-supporting-global-hotkeys.html
in advance thanks
Inspiration webpages
https://blogs.oracle.com/jdevotnharvest/entry/how_to_programmatically_set_focus
http://technology.amis.nl/2008/01/04/adf-11g-rich-faces-focus-on-field-after-button-press-or-ppr-including-javascript-in-ppr-response-and-clientlisteners-client-side-programming-in-adf-faces-rich-client-components-part-2/
how to Commit table by writting Java code in Managed Bean?
Table does not refresh and getting error as UIComponent is Null
A short description of the solution:
(jdeveloper version 11.1.1.2.0)
--- Example where I use onSetFocus in jsff page
<af:commandButton text="#{hrsusuiBundle.FOCUS}" id="cb10"
partialSubmit="true" accessKey="f"
shortDesc="Alt+Shift+F"
actionListener="#{managedBean_clientUtils.onSetFocus}">
<af:clientAttribute name="focusField" value="t1"/>
</af:commandButton>
--- Examples where I use doTableActionAndSetFocus in jsff page
--- There have to be a binding in the jsff page to delete, commit and rollback
<af:commandButton text="#{hrsusuiBundle.DELETE}" id="cb4"
accessKey="x"
shortDesc="Alt+Shift+X"
partialSubmit="true"
actionListener="#{managedBean_clientUtils.doTableActionAndSetFocus}">
<af:clientAttribute name="focusField" value="t1"/>
<af:clientAttribute name="actionField" value="Delete"/>
</af:commandButton>
<af:commandButton text="#{hrsusuiBundle.COMMIT}" id="cb5"
accessKey="s" shortDesc="Alt+Shift+S"
partialSubmit="true"
actionListener="#{managedBean_clientUtils.doTableActionAndSetFocus}">
<af:clientAttribute name="focusField" value="t1"/>
<af:clientAttribute name="actionField" value="Commit"/>
</af:commandButton>
<af:commandButton text="#{hrsusuiBundle.ROLLBACK}" id="cb6"
accessKey="z" shortDesc="Alt+Shift+Z"
partialSubmit="true"
actionListener="#{managedBean_clientUtils.doTableActionAndSetFocus}"
immediate="true">
<af:resetActionListener/>
<af:clientAttribute name="focusField" value="t1"/>
<af:clientAttribute name="actionField" value="Rollback"/>
</af:commandButton>
--- This is the java class I use
--- It is published in adfc-config.xml as a request scope managedbean
public class ClientUtils {
public ClientUtils() {
public void doTableActionAndSetFocus(ActionEvent event) {
RichCommandButton rcb = (RichCommandButton)event.getSource();
String focusOn = (String)rcb.getAttributes().get("focusField");
String actionToDo = (String)rcb.getAttributes().get("actionField");
UIComponent component = null;
String clientId = null;
component = JSFUtils.findComponentInRoot(focusOn);
clientId = component.getClientId(JSFUtils.getFacesContext());
if ( "Delete".equals(actionToDo) || "Commit".equals(actionToDo) || "Rollback".equals(actionToDo) ){
BindingContainer bindings = BindingContext.getCurrent().getCurrentBindingsEntry();
OperationBinding operationBinding = bindings.getOperationBinding(actionToDo);
Object result = operationBinding.execute();
AdfFacesContext.getCurrentInstance().addPartialTarget(component);
if (clientId != null) {
makeSetFocusJavaScript(clientId);
public static String onSetFocus(ActionEvent event) {
RichCommandButton rcb = (RichCommandButton)event.getSource();
String focusOn = (String)rcb.getAttributes().get("focusField");
String clientId = null;
if (focusOn.contains(":")) {
clientId = focusOn;
} else {
clientId = findComponentsClientIdInRoot(focusOn);
if (clientId != null) {
makeSetFocusJavaScript(clientId);
return null;
private static void writeJavaScriptToClient(String script) {
FacesContext fctx = FacesContext.getCurrentInstance();
ExtendedRenderKitService erks = null;
erks = Service.getRenderKitService(fctx, ExtendedRenderKitService.class);
erks.addScript(fctx, script);
public static void makeSetFocusJavaScript(String clientId) {
if (clientId != null) {
StringBuilder script = new StringBuilder();
//use client id to ensure component is found if located in
//naming container
script.append("var textInput = ");
script.append("AdfPage.PAGE.findComponentByAbsoluteId");
script.append ("('"+clientId+"');");
script.append("if(textInput != null){");
script.append("textInput.focus();");
script.append("}");
writeJavaScriptToClient(script.toString());
public static String findComponentsClientIdInRoot(String id) {
UIComponent component = null;
String clientId = null;
component = JSFUtils.findComponentInRoot(id);
clientId = component.getClientId(JSFUtils.getFacesContext());
return clientId;
}

Hi,
I am developing an ADF application where the users do not want to use the mouse. So I would like to know if there are a best practice for this?
Well HTML (and this is the user interface you see) follows a tab index navigation that you follow with "tab" and "shift+tab". Anything else is a short cut for which you use mnemonics (as you already do) or shortcuts (explained in http://one-size-doesnt-fit-all.blogspot.dk/2010/11/adf-ui-shell-supporting-global-hotkeys.html). There is a distinction to make between non-web environments (which I think you and your users have abackground in) and client desktop environments. Browsers block some keyboard functionality for their own purpose. So you may have to find a list of keys first that work across browsers. Unlike desktop clients, which allow you to "press a button" without the button to take focus, this cannot be done on the web. So you need to be clever here, avoiding buttons at all.
The following paper is about JavaScript in ADF and explains the basics for what Chris Muir explains in : http://one-size-doesnt-fit-all.blogspot.dk/2010/11/adf-ui-shell-supporting-global-hotkeys.html
http://www.oracle.com/technetwork/developer-tools/jdev/1-2011-javascript-302460.pdf
It has the outline for how to register short cut keys that perform a specific action (e.g. register ctrl+d to delete the current row you are on, or press F11 to execute a query (similar to Oracle Forms frmres files)). However, be aware that this includes some code you have to write (actually quite some code to be honest).
http://www.oracle.com/technetwork/developer-tools/adf/learnmore/79-global-template-button-strategy-360139.pdf (are there an example implementation?), or
http://one-size-doesnt-fit-all.blogspot.dk/2010/11/adf-ui-shell-supporting-global-hotkeys.html
Actually these are implementations as they come with example code for you to use and customize, do they? So what is this question asking for more ? Also note that global buttons don't quite have anything in common with the question you asked. I assume you want to see it as an implementation of the Forms toolbar that operates on the form or table the focus is in. This however does not work for the web as there is nothing that keeps track of which component has a focus and to what iterator (data block) it belongs. This would involve even more coding (though possibly doable)
Frank

Best Practices for user ORACLE

Hello,
I have few linux servers with user ORACLE.
All the DBAs in the team connecting and working on the servers as ORACLE user and they dont have sperate account.
I create for each DBA its own account and would like them to use it.
The problem is that i dont want to lock the ORACLE account since i need it for installation/upgrade and etc , but yet i dont what
the DBA Team to connect and work with the ORACLE user.
What are the Best Practice for souch case ?
Thanks

To install databases you don't need acces to Oracle.
Also installing 'few databases every month' is fundamentally wrong as your server will run out of resources, and Oracle can host multiple schemas in one database.
"One reason for example is that we have many shell scripts that user ORACLE is the owner of them and only user ORACLE have a privilege to execute them."
Database control in 10g and higher makes 'scripts' obsolete. Also as long as you don't provide w access to the dba group there is nothing wrong in providing x access.
You now have a hybrid situation: they are allowed interactively to screw 'your' databases, yet they aren't allowed to run 'your' script.
Your security 'model' is in urgent need of revision!
Sybrand Bakker
Senior Oracle DBA

Best Practice for Deploying ADF application

I am tasked with developing a best or prefered practice of feploying a large ADF application. Background: we are in the process of redeveloping a UI for a large system. We have broken the system down into susbsytems. Each of these susbsystems UI will be a ADF aaplicaion(?). This is a move from a MS .Net front end. The backend (Batch processes etc) is being dveloped in Java. So my question is if I have several ADF projects for each subsystem and common components that they all will use - what is the best practice to compile package and deploy? The deployment will be to weblogic server or servers(Cluster).
We have a team of at least 40 -50 developers worldwide so we are looking for an automated build and deploy and would like to follow Oracle best practice. So far I have read Deploying ADF Applications (http://download.oracle.com/docs/cd/E15523_01/web.1111/e15470/deploy.htm#BGBJHGFH) and have followed the links. I have also look at the ADF evangalist blogs - lots of chatter about ojdeploy. My concern about ojdeploy is that dependent files are also being compiled at the same time. I expected that we want shared dependent files compiled only once (Is that a valid concern)?
So then when we build the source out of subversion (ojdeploy ? Ant? ) then what is best practice to deploy to a weblogic server (wslt admin console) - again we want it to be automated.
Thank you in advance for replies.
RK

Rule 1: Never use the "Automatically Expose UI Componentes in a New Managed Bean" option, create your bindings manually;
Rule 2: Rule 1 is always right;
Rule 3: In doubts, refer to rule 2.
You may also want to check out :
http://groups.google.com/group/adf-methodology
And :
http://www.oracle.com/technology/products/jdev/collateral/4gl/papers/Introduction_Best_Practices.pdf

Best Practice for thinning Oracle Forms

We are considering upgrading from 10g (10.1.2.3) to (probably) 11g forms initially whist long term migrating to ADF.
Our system is 15 years old and has many issues that go along with a system this old i.e. database keys/indexes missing, forms with lots of code etc.
We have very few resources so thought the best way to go about this would be to thin out all our Oracle forms in 10g prior to upgrade (as this is mainly just a recompile in 11g) but was looking at if there were any best practices as to what level this should be done.
For example we heavily use PLLs. Should the majority of this code be moved into packages or are PLLs ok to use. I take it the majority of program units should be moved to db.
How thin should we aim to make the form. i.e if we are using post queries for something as simple as select customername into :nondbitem from customer where customerid =:customerid. Should we be making this a function in the db and reusing the code in the necessary places?
Just wondering what people would deam best practice before we start.
Thanks

Thanks for all your input.
Although the current database design is not correct we are going to correct the existng database rather than rewriting it.
So were hoping to correct this also!
But what im doing now to get rid some of this post queries for nondbnames is im creating a view of all information i need to display on one pertaining record.
Then i used this view like block basic procedure like relationship to the transactional block.So if you make queries or insert (need to run query the block view) it will give all the necessary record display"
{CODE}
This sounds sensible but we havs hundreds of tables with lots of columns. How are you managing the views/deciding how manytables/columns to include on the forms?

User Authorization, best practices for this custom application requirement?

JDeveloper 12c (12.1.2)
We want to use external LDAP (active directory) with ADF security to authenticate and authorize users.
One of our custom application requirements is that there is a single page with many interactive components. It has probably about 15 tables and each table would need to have the following buttons (or similar components):
- delete: (if certain row is selected) to delete it
- edit: (if certain row is selected) takes user to 'edit page' where changes can be made
- create: to create new record for this particular VO (table)
So let's say that would be 3 x 15 = 45 different actions that single user can possibly perform. Not all users have same 'powers' ie some users can only edit CERTAIN tables, and delete from one or two. Most users can create and edit most VOs etc
Back when this application was originally developed using (I believe) 10g JDeveloper with UIX, the way it was done is that we maintained a table in database with 'user credentials' as Y or N flags.
For example: DEL_VO1, EDIT_VO1, ADD_VO1....
So when user is authenticated we would then pull all these credentials from the DB table and load them into the session variables. Then we would use EL to render or not render certain buttons on the page. For example: rendered="#{sessionScope.appDelVo1 == 'Y'}"
Moving forward into latest ADF technology, what would be the best practice to achieve described functionality?

Hi,
ADF BC could have permissions added to the entity level (includes remove and update). So you can create permissions for the entity (as it doesn't matter for data security how data is accessed. If as a user you are nit allowed to change a database table then this is for tables and forms). You can then use EL to check the permission, thus no need to keep the privileges in the database.
If a user is allowed to update an entity then you can check this using EL in the UI
<af:inputText value="#{bindings.DepartmentName.inputValue}"
readOnly="#{!bindings.DepartmentName.hints.updateable}">
whatch this for a full coverage of ADF Security: Oracle ADF Security Overview - Oracle JDeveloper 11g R1 and R2
Frank

Best practices for a large application

I have an existing application that is written in character
based Oracle Developer and has many forms and reports. A few years
ago I converted an inquiry portion of the system to standard ASP,
this portion has about 25 pages. Initially I want to rewrite the
Inquiry portion in Flex (partly to teach myself flex), but
eventually (soon) I will need to convert the remainder of the
application, so I want a flexible and robust framework from the
beginning.
So far for fun I wrote a simple query and I like what I have
done but I realized that trying to write the entire application in
a single script with hundreds of states would be impossible. The
application is a fairly traditional type app with a login script, a
horizontal menu bar and panels that allow data entry, queries and
reports. In Oracle and ASP each "panel" is a seperate program and
is loaded as needed. I see advantages in this approach, and would
like to continue it (creating as much reusable stuff as possible).
So where can I find documentation, and or examples on the
best practice in laying out a framework for what will eventually be
a sizeable app?
BTW, what about the ever present problem of writing reports?
Paul

As a matter of fact, modules are exactly what you're looking
for! :)
Flex's doc team has a good intro document here:
http://blogs.adobe.com/flexdoc/2007/01/modules_documentation_update.html
And this gentleman has a quick example in his blog:
http://blog.flexexamples.com/2007/08/06/building-a-simple-flex-module/
The app I'm working on now is primarily based on modules, and
they're a good way to keep things organized and keep loading times
low. Documentation on them has been so-so, but its getting better.
There are little undocumented gotchas that you'll undoubtedly run
into as you develop, but they've been covered here and in the
flexcoders Yahoo group time and again, so you'll readily be able to
find help.
When creating a module, you can either place them in the
project that they'll eventually be used with, or create a stand
alone project just for that module (the preferred and more
organized method.) Creating a stand-alone module project is easy:
just change the "mx:Application" tag in your main mxml file to a
"mx:Module" tag instead.
If you're using Cairngorm as your app's framework, you'll
have to tinker with things a bit to get it all working smoothly,
but its not that hard, and I believe the doc team is working on a
definitive method for using modules in Cairngorm based apps.
Hope this helps!

Best practices for administering Oracle Big Data Appliance

-        Best practices as part of administration of Oracle Big Data Infrastructure
-        How do we lock down max space usage per project
Eg: Project team A can have a max limit of 10 TB space allocated
-        Restricting roles, access ( Read, Write), place holder for common shared artifacts
-        Template/procedure for code migration across dev,qa and prod environments etc

Your data is bigger than I run, but what I have done in the past is to restrict their accounts to a separate datafile and limit its size to the max that I want for them to use: create objects restricted to accommodate the location.

What are best practices for rolling out cluster upgrade?

Hello,
I am looking for your input on the approaches to implement Production RAC upgrade without having a spare instance of RAC servers. We have a 2-node database RAC 11.1 on Production. We are planning to upgrade to 11.2 but only have a single database that the pre-Production integration can be verified on. Our concern is that the integration may behave differently on RAC vs. the signle database instance. How everybody else approaches this problem?

you want to test RAC upgrade on NON RAC database. If you ask me that is a risk but it depends on may things
Application configuration - If your application is configured for RAC, FAN etc. you cannot test it on non RAC systems
Cluster upgrade - If your standalone database is RAC one node you can probably test your cluster upgrade there. If you have non RAC database then you will not be able to test cluster upgrade or CRS
Database upgrade - There are differences when you upgrade RAC vs non RAC database which you will not be able to test
I think the best way for you is to convert your standalone database to RAC one node database and test it. that will take you close to multi node RAC

Best Practices for 11i to R12 Upgrade

Hi,
We are in the process of doing an upgrade from 11.5.7 to R12.
Please share the Best Practices available for upgrade from 11.5.7 to R12.
Regards,
Narayan

Hi Narayan;
Please see:
Interesting Documents Concerning E-Business Suite 11i to R12 Upgrades [ID 850008.1]
Upgrade Advisor: E-Business Suite (EBS) Technology Stack Upgrade from 11.5.10.2 to 12.1.2 [ID 253.1]
Those should helps you
I suggest also use search mechanisim and make search on EBS forum, you can see our previous topic which is mention same&similar issues
Regard
Helios

Best practices for cleaning up and upgrading

Apologies in advance, as I know there is a wealth of info and ideas on this topic, but I'm looking for the best possible combination of simplicity and effectiveness and the community has done well by me in the past.
Situation: I have this old MacBook
Model Name: MacBook
Model Identifier: MacBook6,1
Processor Name: Intel Core 2 Duo
Processor Speed: 2.26 GHz
Number of Processors: 1
Total Number of Cores: 2
L2 Cache: 3 MB
Memory: 4 GB
Bus Speed: 1.07 GHz
Boot ROM Version: MB61.00C8.B00
System Version: OS X 10.9.5 (13F34)
Kernel Version: Darwin 13.4.0
HD- Available: 6.22 GB
Capacity: 249.2 GB
Literally the only paid software that I have on here and need to keep is Microsoft Office. I do not have the original install disk for this software.
This machine is running very very slowly, and not much hard disk space left. I'm not doing very well at cleaning up.
I have brand new Time Machine backup and I have some pretty old Time Machine backups, pre Mavericks.
I have a new 2TB external drive.
I'm looking for the most effective way to speed this one up and make it as lean as possible and, if advisable, upgrade to Yosemite.
Any recommendations?
I'm willing to spend a little money if it can make a substantial difference.
Thanks and apologies for length of post.
RW

Hi,
First thanks for this. This application has been really helpful. I've deleted a lot of the easy stuff (music, pics, videos) and am now getting down to some of the stuff I'm less comfortable whacking without some guidance. Here's an example. I don't use the Mail application at all (use Gmail). Here is what I see on OmniDisk Sweeper:
https://www.dropbox.com/s/6iedkj46u4jaxbf/Screenshot%202014-11-08%2010.26.33.png ?dl=0
I think the bulk of this is just messages from my gmail account, from a failed attempt to use Mail a bit years ago. So can you help with at which point I can delete? Hope question make sense. You've helped a lot so far and I've freed up about 45 gb so far.
Rob

Best practice for Java stand alone upgrade into maint org area?

I have a Java standalone running EP 6.0 on SAP netweaver 04 at SP 20.
I want to upgrade to SAP Netweaver 2004s. I think the equivalent of SPS 20 of NW 04 in NW 2004s is SPS 11 (ie all the SCAs are named *11...).
NW 2004s SPS 11 download requires use of maintenance_organizer.
Does this mean I should change the SMSY definition of my java solution to reflect SAP Netweaver 2004s now, rather than after I actually upgrade, even though the upgrade may be several weeks away?
Is changing the SMSY def the only way to get the SPS 11 media?
I am currently in the process of collecting all of the media required for a sucessful upgrade.
Thx
Ken Chamberlain
University of Toronto

kevjava wrote: Some things that I think would be useful:
Suggestions reordered to suit my reply..
kevjava wrote: 2. Line numbering, and/or a line counter so you can see how much scrolling you're going to be imposing on the forum readers.
Good idea, and since the line count is only a handful of lines of code to implement, I took that option. See the [line count|http://pscode.org/stbc/help.html#linecount] section of the (new) [STBC Help|http://pscode.org/stbc/help.html] page for more details. (Insert plaintiff whining about the arbitrary limits set - here).
I considered adding line length checking, but the [Text Width Checker|http://pscode.org/twc/] ('sold separately') already has that covered, and I would prefer to keep this tool more specific to compilation, which leads me to..
kevjava wrote: 1. A button to run the code, to see that it demonstrates the problem that you wish for the forum to solve...
Interesting idea, but I think that is better suited to a more full blown (but still relatively simple) GUId compiler. I am not fully decided that running a class is unsuited to STBC, but I am more likely to implement a clickable list of compilation errors, than a 'run' button.
On the other hand I am thinking the clickable error list is also better suited to an altogether more abled compiler, so don't hold your breath to see either in the STBC.
You might note I have not bothered to update the screenshots to show the line count label. That is because I am still considering error lists and running code, and open to further suggestion (not because I am just slack!). If the screenshots update to include the line count but nothing else, take that as a sign. ;-)
Thanks for your ideas. The line count alone is worth a few Dukes.

Best practices for an oracle application upgrade

Similar Messages

Maybe you are looking for