Exploded vs. Archived EAR structure (when using split development)

Using the split development structure suggested in Weblogic 8.1 with wlpackage
ant task, I noticed that the exploded ear and the archived ear generation creates
a "flat" deployment structure. i.e. all the files inside the ear hierarchy are
bundled without any intermediate respective jars.
Is there a provision to package the ejb modules in one or more jar files when
building a qa and/or production ear ? Is this possible to accomplish or do I need
to write separate ant "targets" in the ant script that would manually jar up the
necessary ejbs. If so, how would this work in conjunction when I have a "ant ear"
target ?
I also noticed that the medrec example app has application specific value/dto
classes bundled in a separate jar file using a build.xml file. My question, when
running wlcompile at a higher level how does it know NOT to put the value classes
separately under APP-INF/classes, i.e. why does the build not build/see these
common value classes twice.
Thanks for your help!

DP wrote:
Using the split development structure suggested in Weblogic 8.1 with wlpackage
ant task, I noticed that the exploded ear and the archived ear generation creates
a "flat" deployment structure. i.e. all the files inside the ear hierarchy are
bundled without any intermediate respective jars.
Is there a provision to package the ejb modules in one or more jar files when
building a qa and/or production ear ? Is this possible to accomplish or do I need
to write separate ant "targets" in the ant script that would manually jar up the
necessary ejbs. If so, how would this work in conjunction when I have a "ant ear"
target ? You would have to do this manually. Typically wlpackage would have to
modify your application.xml to archive the modules. For instance, your
application.xml probably has uris like myweb, myejb in it now because
that's how you named your directories. If we archived those, they'd
become myweb.war and myejb.jar, and the application.xml would need to be
modified to reflect the new uris.
It's all possible for us to do, but we ran out of time during 8.1 to get
this one in.
>
I also noticed that the medrec example app has application specific value/dto
classes bundled in a separate jar file using a build.xml file. My question, when
running wlcompile at a higher level how does it know NOT to put the value classes
separately under APP-INF/classes, i.e. why does the build not build/see these
common value classes twice. I'm not sure I exactly understand your question, but let me explain how
wlcompile works and see if that answers your question.
wlcompile looks at each directory under the srcdir and decides whether
it's an EJB, a webapp, or a java module. It supports an
includes/excludes list to let you filter those directories.
Anything that's not recognized as an EJB or webapp is compiled into
APP-INF/classes.
-- Rob
>
Thanks for your help!

Similar Messages

VPRS incorrect cost when using Split Valuation

Hello,
We are using split valuation thus each valuation type has a separate moving average price. When creating a billing doc, my pricing procedure is called which includes VPRS. The problem is the cost returned is not for the correct valuation type which is passed in the billing document (Field BWTAR). The cost returned in pricing is the moving average without any valuation type. Does anyone know of any configuration to calculate VPRS based on the batch/valuation type provided in the billing document?
My pricing procedure includes condition VPRS which is set to SUTOT = B, Reqt = 4 and statistical. The condition type is set to defaults (Cond Class = B, Calculation Type = C, Condition Category = G). The article master has field Price Control set to V and split valuation is enabled.
The billing document type being used is FP (POS Interface)
The item category is DLN
Any help would be greatly appreciated.
Thanks,
Pierre

Hello Pierre
I am not totally knowledgeable on this subject but want to give my best shot.
I think you need to ensure that the system takes (redetermines) the cost from Valuation segment in the material master.
1) So make the cond category a T (Moving price). Don't change VPRS but copy it and set up ZPRS and then use it in the pricing procedure as an alternative to VPRS.
2) You may also need to work on the copy control- Item cat level- especially Pricing Rules- Probably you need to set up a custom one which says 'keep manual prices as they are, but redetermine cost". Then you need to set up USEREXIT_PRICING_RULE from Program RV61AFZA to make this work. . Use OSS note 24832 to set up your new pricing rule.
3) Also check out 'Price Source' field- in copy control again- to see if it helps you achieve your objective.
Also about VPRS, OSS note 547570 says "If the condition control (KSTEU) is set to 'H', the cost was taken from the goods issue. If it set to 'A', it was redetermined from the valuation segment of the material master, in case of 'D' or 'E' it was copied from the preceding document:. But I don't know how you can use this information.
Go through these OSS notes and those mentioned within for a good idea of VPRS.
372772 - How does the system determine the cost?
547570 - FAQ: VPRS in pricing
24832 - Pricing rules / TVCPF
371844 - Correction of incorrect costs
372760 - Several problems with VPRS
Hope this helps, Let me know.

Connection Problem When Using Oracle Developer Tools for Visual Studio

I tried to create an Oracle Connection in my Visual Studio 2008 after installing Oracle 11g, Oracle Data Access Components (ODAC) with Oracle Devleoper Tools for Vissual Studio version 11.1.0.6.20.
I followed the directions according to "Building .NET Applications Using Oracle Developer Tools for Visual Studio, when I click Test Connection, I keep getting the error message below.
Microsoft Visual Studio
ORA-12170: TNS:Connect timeout occurred
OK
---------------------------

I solved the problem.
This is what I did for the benefit of the embryonic newbe like me.
On the Add Connection Dialog
1.     For Data source name I selected Local Database
2.     I clicked on the specific User Name and I typed SYSTEM
3.     For Password, I typed orcl (the password I setup during the installation) (HR did not work for me for User name and Password according to the tutorial instructions)
4.     For Role, I selected Default
5.     Connection Name, I selected Local Database
6.     Before proceeding any further went to Windows XP SP3 where I was operating from, under Start Menu, I selected Administrative Tools then Services
a.     On the Windows Services (Local) Dialog, under name, I selected OracleServiceORCL and OracleOraDb11g_home1TNSListener. I clicked on them one at a time
b.     On the presented Dialog, under Service Status, I Clicked on the Start Button to start the Services
c.     Then I went back to the Add Connection Dialog. I clicked on Test Connection, it connected okay. I finally clicked on Okay to connect.

Develop/Deploy Adapters/Connectors using split development directories

BEA documentation says that 8.1 does not support resource adapter split development/deployment
for writing adapters from scratch. We will be using split directories and writing
connectors for these. What are our options ? (we would still like to develop and
deploy ejbs and web applications using the split dev/deployment).

wlcompile won't recognize them or do anything special. It would just
compile them into APP-INF/classes. You could just exclude that module
and compile it yourself if you wanted more control over it.
wlappc works fine with them, and split-directory should work fine as well.
-- Rob
DP wrote:
BEA documentation says that 8.1 does not support resource adapter split development/deployment
for writing adapters from scratch. We will be using split directories and writing
connectors for these. What are our options ? (we would still like to develop and
deploy ejbs and web applications using the split dev/deployment).

Error when using sql developer

my computer lost power while i was using sql developer (version 1.1.0.23). i restarted and logged in and now any time i try to access a saved db connection or go to tools prefernces, database and any parameters i get the following error or something very similar.
ava.lang.NullPointerException
     at oracle.dbtools.raptor.conntypes.RaptorClassLoader.findClassLoader(RaptorClassLoader.java:48)
     at oracle.dbtools.raptor.conntypes.RaptorConnectionWrapper.createConnectionImpl(RaptorConnectionWrapper.java:202)
     at oracle.dbtools.raptor.conntypes.RaptorConnectionWrapper.createConnection(RaptorConnectionWrapper.java:160)
     at oracle.dbtools.raptor.conntypes.RaptorConnectionWrapper.getJdbcConnection(RaptorConnectionWrapper.java:191)
     at oracle.dbtools.raptor.conntypes.RaptorConnectionWrapper.getPresentation(RaptorConnectionWrapper.java:66)
     at oracle.dbtools.raptor.utils.Connections$ConnectionInfo.createConnection(Connections.java:738)
     at oracle.dbtools.raptor.utils.Connections$ConnectionInfo$ConnectRunnable.doWork(Connections.java:611)
     at oracle.ide.dialogs.ProgressRunnable.run(ProgressRunnable.java:159)
     at oracle.ide.dialogs.ProgressBar.run(ProgressBar.java:553)
     at java.lang.Thread.run(Thread.java:619)

Download the latest and Greatest 3.1 at :http://www.oracle.com/technetwork/developer-tools/sql-developer/overview/index.html
1.1 is a very old version
Michael

Nested structure when using FM XXL_SIMPLE_API

I use this logic to fill in the headings before calling the FM xxl_simple_api in one of my programs. Is it possible to have nested structure(for <b>i_tab</b> in my code) and still use same kind of logic to fill in the headings? Not sure if this Fm would display data correctly when there is a component of strucutre type. I tried it, the field was blank with no contents in excel.
DATA : nlines TYPE i,
         tempfile LIKE gxxlt_f-file,
         descr_ref TYPE REF TO cl_abap_structdescr,
         header LIKE gxxlt_v OCCURS 0 WITH HEADER LINE,
         it_print LIKE gxxlt_p OCCURS 0 WITH HEADER LINE.
*data:               keycol type i.
FIELD-SYMBOLS:
<comp_wa> TYPE abap_compdescr,
tempfile = file_name.
REFRESH header.
CLEAR header.
descr_ref ?= cl_abap_typedescr=>describe_by_data( i_tab ).
DESCRIBE TABLE descr_ref->components LINES nlines.
nlines = nlines + 1.
LOOP AT descr_ref->components ASSIGNING <comp_wa>.
    IF sy-tabix = nlines.
      EXIT.
    ELSE.
      header-col_no = sy-tabix.
      header-col_name = <comp_wa>-name.
      APPEND header.
    ENDIF.
ENDLOOP.
CALL FUNCTION 'XXL_SIMPLE_API'
EXPORTING
    filename = tempfile
N_KEY_COLS = keycol
TABLES
   col_text = header[]
I appreciate your help. Thanks

Hi,
It's better to use another FM 'SAP_CONVERT_TO_XLS_FORMAT'
call function 'SAP_CONVERT_TO_XLS_FORMAT'
             exporting
              I_FIELD_SEPERATOR = seper
              I_LINE_HEADER = 'X'
                 I_FILENAME = P_FILE
              tables
                  I_TAB_SAP_DATA = itab
              exceptions
                  CONVERSION_FAILED = 1.
Svetlin

Issue with OM relationship structure when using report RHSTRU00

In HR, one Person (P) is attached to only one Central Person (CP). So, when we run the report with date period as of Today, we expect a Tree structure in the output with P-CP relationships displayed. However, we see that the report displays one Person attached to many CPs.
The program fetches the 1 P to 1 CP combination. However, it deletes all Persons who are not Active as of today but does not delete the CP. It then attaches all CPs to the first Active Person it can find.
This is a serious bug since it shows data which is different from what exists in the system. In our systems as would be in any HR system in the world, One person can be attached to only one CP and the program shows that the Person is attached to many CPs.
I opened a OSS message but SAP refused to correct this bug.
Has anybody else ever encountered the same issue and do you have a solution to this?

I checked the report, and it certainly returns incorrect data, I found exactly what you say regarding active/inactive persons.
What is SAP's reason for not looking into this? It would be interesting to hear their argumentation for this not being a bug.
It may be that SAP expects RHSTRU00 to be called only for active persons? This would require running another report first to get the input for RHSTRU00. This seems like a strange prerequisite for the program, and I do not understand why SAP don't see this as a bug. The error seems rather obvious and reproducable.
I often find that I have to push to get passed first line of support to have SAP see the problem. It is tiring, but I have found this to be the fact far too often.
PS - we never use the program like this, so we do not face the problem ourselves.
/Kirsten

When can I use my WSABUF and overlapped structures when using WSASendTo on an overlapped socket

When calling WSASendto on a UDP socket that is overlapped it is difficult to get the right understanding from the documentation provided in MSDN.
The question is, will WSASento copy the content of my WSABUF's so I can use them again? (Even before the complete routine and the event is signalled)
According to http://windowssdk.msdn.microsoft.com/en-us/library/ms741693.aspx this is the case:
"The array of WSABUF structures indicated by the lpBuffers parameter is transient. If this operation is completed in an overlapped manner, it is the service provider must capture these WSABUF structures before returning from this call. This enables applications to build stack-based WSABUF arrays."
But reading a Microsoft book in chapter 6 conserning Winsock and QOS:
"On the sending side, applications use WSASend or WSASendTO to supply pointers to filled buffers and then agree not to disturb the buffers in any way until the network has consumed the buffer's contents"
Can I call WSASendTo again before completion of the last WSASenTo is indicated ?
Do I have to give a new WSABUF every time ? Or can I reuse it ?
Can I reuse the Overlapped structur as well ? Or do I need to wait for the completion for the last WSASendTo ?
Background information:
I am doing an application that is quite time critical. I use one sending UDP socket and a receiving UDP socket. Both of them are asyncron. The receiving socket is working fine. The reason for chosing asyncron (overlapped) for the sending socket is because sometimes a WSAsendto can last 3,5 second if the ip that I wish to send to is unknown to to PC. But I need to send the next UDP packet in around 1msek. In normal conditions it only takes around 50usek to calls WSASendto.
I hope somebody can help me
Best regards
Flemming

Flemming BT wrote:
Thanks for the answer. But I will like to dig a little bit in it to be absolutely sure.
The chapter 6 of the Microsoft book can be found at http://www.thaifast.com/Microsoft_Books/Sample/02783006.pdf
This chapter is dealing with Winsock 2 and the quote is from page 12 (or 74 according to print) is about overlapped I/O.
I am considering code like this (note that the overlapped is a menber and initiated once in the constructor, the menber function this is called in can be called before the actual send is finished(that is before the event on the overlapped structure is signalled), can the system handle this ?). Actually if the UDP message is send or not is not interesting for me, because timeouts on the respons on the message are handled somewhere else in my program, the reason for this is that I use UDP and the packets could get lost somewhere in the network anyway:
                WSABUF wsaBuffer;
                LPWSABUF lpBuffers;
                lpBuffers = &wsaBuffer ;
                wsaBuffer.len = nBufLen;
                char *pTempBuf = new char[nBufLen];
                for (int i = 0; i < nBufLen;i++)
                        pTempBuf = lpBuf;
                wsaBuffer.buf = pTempBuf;
                DWORD noOfBytesSend = 0xffffffff;
                int socketSize = sizeof(sockAddr);
                nResult = WSASendTo(m_hSocket,lpBuffers,1,&noOfBytesSend,0,(SOCKADDR*)&sockAddr,socketSize,&m_overLappedSend,,NULL);
                if (nResult != 0)
                        int error = WSAGetLastError();
            if (error == WSA_IO_PENDING)
                                TRACE("Output is pending\n");
                        TRACE("SendTo returned %d error %d\n",nResult,error);
                delete [] pTempBuf;
                pTempBuf = NULL;
I'd put my money on the MSDN documentation over a single book. The documentation for WSASendTo has been around, scrutinized, and edit for years. If it says you can use stack memory in a call to WSASendTo that would mean that you're free to do whatever you want with the memory given to WSASentTo as soon as WSASendTo returns.
I'll see if I can dig up an exact (rather than inferring from the documentation) what's really going on with WSASendTo.

Error when using SQL Developer to IMPORT Access data into Oracle

Hello,
I am using SQL Develoepr to export Access mdb file into Oracle. When I try to use Tools ---? Migration---> Microsssfot access exported -->
I am getting an error message upon attaching the Access database file, the error states
"Error#40179 - XMLExporter - Methiod'AddFromFile' of object'_References failed. Export did not complete successfully.
What could be the reason for this error please advise.
Thanks

Hi,
Try searching for:Re: 30EA2 - No Available Databases in Capture step with MS Access Migration
Thread: 30EA2 - No Available Databases in Capture step with MS Access Migration
Posted: Dec 20, 2010 11:04 AM
Also note that copy to oracle - (for just table/column information) is available on right click.
-Turloch
SQLDeveloper team

Unable to COMMIT when using SQL Developer 4

Hello guys,
I was wondering if any of you have an issue like mine? Whenever I try to commit to SVN, I get the following exception:
{quote} svn: E170001: Commit failed (details follow): svn: E170001: HTTP proxy authorization failed {quote}
Thanks,
{code}
Java(TM) Platform
1.7.0_25
Oracle IDE
4.0.0.13.80
Versioning Support
4.0.0.13.80
{code}

It sounds like network problem (such of firewall, NAT...) Can you open a sqlplus connection? Fine work.
Please check in this order, to localize the problem
ping <to_ip_machine>
tnsping <connnect_string>
sqlplus <connect_string>
HTH - Antonio NAVARRO
sql1.wordpress.com

Developing using EAR structure gives redeploy/classpath issues

I am very keen to get people's opinions and methods of developing and packaging
EARs/WARs/EJBs under WL6.1 sp2.
What method do you use and what problems do you get...
We have recently migrated to WL6.1 sp2, and looked forward to using the Enterprise
Applications Archive expanded structure to deploy and develop under. However I
have come across a few fundamental problems that seem to make development in this
structure worthless.
It seems it's easier just have all the expanded classes paths on the Java Classpath
and just restart the server after a individual class file recompile ?
Problems :-
- With an expanded EAR structure, I believe you can only redeploy the WHOLE EAR
(by touching the REDEPLOY file), not an individual ejb jar or just the WAR part.
I have tried putting a REDEPLOY file in the WAR \WEB-INF\ directory, as recommended,
but I think this only works outside of the EAR structure. If you use the Console
to just undeploy an redeploy the Web App, it redeploys the whole EAR. So this
leaves us having to undeploy everything to test out a change.
- The classloaders hierarchy available with an EAR structure seems ideal at first,
e.g. the Web App can see all EJB jar classes. We have our EJB jars containing
the beans and just the DEPENDANT classes (often these are abstract types), this
keeps the jars lean and avoids duplicated classes across ejb jars. But this leaves
some classes which are not in any EJB jar at all (often subclasses or implementations
of these abstract types), however the Web App still needs them. It is difficult
to isolate these individual classes. So for now we have two solutions which don't
seem quite right :-
Make one of the EJB.jars contain ALL the Server side classes, not just dependant
classes. Or make a full jar of ALL server side classes and add this jar to the
Web Apps \META-INF\manifest Class-Path like you would do with an EJB jar.
Is there a way to have a non-EJB jar at the EJB classloader level ? This would
solve this and allow Web Apps to see other non-ejb jars.
How do you guys works and avoid the problems we've hit ?

Pete <[email protected]> wrote:
Is there a way to have a non-EJB jar at the EJB classloader level ? Sure, just add it to the Class-Path: manifest entry in your ejb.jar.
Dimitri

How to archive an exploded format into EAR

Hi all,
I have my application in the following exploded format:
c:\bea\...\mydomain\applications\MyApplication
MyApplication
\ejb
\mypackage\....
\META-INF\ejb-jar.xml, weblogic-...........
\META-INF\application.cml, weblogic-application.xml
\web
\folder1\*.jsp
\folder2\*.jsp
\WEB-INF\web.xml
\WEB-INF\lib\*.jar
\WEB-INF\tld\*.tld
\WEB-INF\classes\mypackage1\*.java
Now, anyone please tell me how to package the above exploded format into .jar
and war and finally the EAR.
and also, after deploying the EAR, how can we access the application?
Thanks in Advance,
Seshu

The bottom line is you should use ant.
If your using a version prior to 8.1 you can use the built in jar/ear tasks,
if you cant upgrade to 8.1 I reccomend it.
If your using 8.1 we have an exploded directory structure for development
known as split-dir (go to http://edocs.bea.com and search) with a set of ant
tasks to do compilation and deployment. There is also one called wlpackage
that will create an archived EAR for you as well.
The basic principle is that in development you want to develop with an
exploded format as it is much faster to compile and to deploy. When moving
to production and for distribution you want an archive. You have a couple of
options for the archive. You can archive EVERYTHING where each module
inside of the ear is also archived.
Such as:
foo.ear
META-INF/applicaiton.xml
web.war
ejb.jar
I am not a huge fan of this because in development you really want
everything exploded and if you archive your modules inside the ear you have
to change the URI's in your applicaiton.xml. I prefer:
foo.ear
META-INF/applicaiton.xml
web/
WEB-INF/web.xml
ejb/
META-INF/ejb-jar.xml
This is the format of the ear that wlpacage ant task creates for you.
Cheers
mbg
"Seshagiri" <[email protected]> wrote in message
news:3f08c9ea$[email protected]..
>
Hi all,
I have my application in the following exploded format:
c:\bea\...\mydomain\applications\MyApplication
MyApplication
\ejb
\mypackage\....
\META-INF\ejb-jar.xml, weblogic-...........
\META-INF\application.cml, weblogic-application.xml
\web
\folder1\*.jsp
\folder2\*.jsp
\WEB-INF\web.xml
\WEB-INF\lib\*.jar
\WEB-INF\tld\*.tld
\WEB-INF\classes\mypackage1\*.java
Now, anyone please tell me how to package the above exploded format into.jar
and war and finally the EAR.
and also, after deploying the EAR, how can we access the application?
Thanks in Advance,
Seshu

Exploded Vs Archive form of EAR file

Hi
I have 3 JSPs named one.jsp, two.jsp, three.jsp.
one.jsp includes two.jsp in 4 places.
two.jsp includes three.jsp in 3 places.
All the three jsps uses JSTL to load the values from sql server database.
When I load the one.jsp file from an Archived ear file, it is taking 7 to 10 seconds to load the file. But, When I try to load it from an exploded ear file, it is loading within a second.
What is the difference between loading the same file in exploded and archived forms?
what are the adavantages and disadvantages of Exploded Vs Archive form of an web application?
Please, help to know this information.
Thanks,
JavaBuddy.

When I load the one.jsp file from an Archived ear
file, it is taking 7 to 10 seconds to load the file.
But, When I try to load it from an exploded ear file,
it is loading within a second.Could you clarify what do you mean by load the file? Do you mean redeploy the application?
And how did you deploy the web application? Through autodeploy, cli deploy or admin console, netbeans?
Do you have precompilejsp option set to same value for both archive and directory deployment (what you referred as the exploded)? This option is used to dermine whether we compile jsp files during deployment time or runtime when the jsp is invoked the first time..
>
What is the difference between loading the same file
in exploded and archived forms?
what are the adavantages and disadvantages of
Exploded Vs Archive form of an web application?Directory deployment has less overhead than archive deployment (no need to explode the archive etc) so it's usually faster.
Directory deployment is targeting developers who need to change the contents of their application often and redeploy. It is faster and also developer doesn't need to repackage the archive each time after modifying the contents.
Directory deployment has some limitations: currently it is not supported on EE platform (cluster). If you don't need to use the cluster feature and always use the PE version, there should be no problem just always doing directory deployment.

How to use the "split development directory structure" in JBuilder9 ?

How to use the "split development directory structure" in JBuilder9 ?
Have /APP-INF supported ?

I've not used JBuilder much, but if it has ant support, then it should
be pretty straight-forward.
-- Rob
Rick Ni wrote:
How to use the "split development directory structure" in JBuilder9 ?
Have /APP-INF supported ?

HT3310 Why can't I hear when using my ear plugs on my I-pad

WHy can't I hear when using my apple ear phones on my I-pad?

Make sure that they are inserted all the way into the jack. It takes a firm push to insert them all the way.

Exploded vs. Archived EAR structure (when using split development)

Similar Messages

Maybe you are looking for