Perfomance problems changingTransaction Attribute

Hi,
I am working with weblogic 7.0. All my EJB's Transaction Attribute value is defined
as "Required". I would like to change this attribute's value to some of my query
methods to "Never" or "Not Supported" value.
Should this change improve the perfomance of these methods?.
In the tests i have made instead of improving performance it has decreased.
I know that in weblogic Server 7.0 exists an optimization to improve the performance
of Remote Method Interface calls within the servers, which permits calling methods
using pass by reference.
Could this performance reduction previosly mentioned, by changing Transaction
Attribute value to "Never" or "Not Supported" to methods, be related with this
optimization included in weblogic ?
Is this optimization available only when this attribute is set to "required"?
Thank's

Hi Davis,
"David Catalán" <[email protected]> wrote in message
news:3eedd1d5$[email protected]..
I am working with weblogic 7.0. All my EJB's Transaction Attribute valueis defined
as "Required". I would like to change this attribute's value to some of myquery
methods to "Never" or "Not Supported" value.
Should this change improve the performance of these methods?.No, it will not, al least for database operations, as it will be
equivalent an autocommit mode which is more expensive as
the database will have to do rapid fire one-statement transactions.
In addition, it will involve container overhead connected with
necessity to suspend and resume ongoing TXs.
In the tests i have made instead of improving performance it hasdecreased.
No surprise.
I know that in weblogic Server 7.0 exists an optimization to improve theperformance
of Remote Method Interface calls within the servers, which permits callingmethods
using pass by reference.
Could this performance reduction previosly mentioned, by changingTransaction
Attribute value to "Never" or "Not Supported" to methods, be related withthis
optimization included in weblogic ?No, the remote call optimization is done without regard
to transaction attributes.
Is this optimization available only when this attribute is set to"required"?
No, it's not limited to "Required". It's available whenever an EJB and a
client
are packaged into one jar or ear.
HTH.
Regards,
Slava Imeshev

Similar Messages

Fetching materials on basis of Division creating perfomance problem

Hi Experts,
i am fetching Data from three table like
select a~matnr a~extwg
b~charg b~werks b~clabs b~cspem b~cinsm
c~vfdat
from mara as a join mchb as b
on a~matnr = b~matnr join mch1 as c on b~matnr = c~matnr and b~charg = c~charg
into table itab_main
where a~spart = p_spart and b~werks = p_werks
and b~lgort = 'FG01' .
it is creating perfomance problem
so my questin is 'is there any index table for mara ?(like vrpma for vbak)'
so we can get all the material on base of division fast.
Regards,
Alpesh
Edited by: Alpesh on Jun 2, 2009 3:56 PM

Hi Alpesh,
How about throwing MARC into the mix. I notice that you are searching with plant, storage location and division. Table MARC has an index on plant named WRK.
If you join MARA and MARC to first determine all the materials that qualify for the given plant and division combination, you can access table MCHB with material (MARA-MATNR), plant (MARC-WERKS) and storage location (FG01) i.e. the first 4 fields in the primary key of table MCHB.
I could not test this theory out because in our system we do not have any data in table MCHB. Your select executes quickly on my system.
Please let me know how this panned out.

Basic Problems with Attributes and Data Loading

Hi Specialists,
I have a basic understanding Problem. I will explain it with a simple Example. I have to load Employee-Data from a file. For each Employee we have
EmployeeNo
EmployeeName
EmployeeLanguage
My Modelling:
1 InfoObject Employee, with EmployeeNo as Key and EmployeeName as Text. For the EmployeeLanguage I create an additional InfoObject, in order to have an Attribute EmployeeLanguage at the InfoObject Employee. So far so good (I think).
What kind of DataSources have I to create?? It is necessary to load the possible Languages first?? If yes, what kind of DS I have to choose, for loading only "English", "German", "Dutch". Attribute? Text?
If not, the loading process is not possible, because one language exists more than one time in my flat file.
Hope you can help me to see it clear.
Thanks.
Denise

Hi Denise,
I don't think you have to load all possible languages first. This is only necessary if you also want to load texts for EmployeeLanguage (example language german with texts "german", "deutsch", ...).
Specify Employee as data target (in tab Master Data / Texts) and assign your InfoArea. There you will find your targets for texts and attributes, now create InfoSource for your file and assign update rules for both texts and attributes.
Best regards,
Björn

Problem with attributes labels

Hi All!
I try use Busines Componets properties. Exactly attributes labels. I setup labels for each attribute, but no effect on rendering. Anybody have this problem?
I use JD 3.1

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Mike Evteev ([email protected]):
Hi All!
I try use Busines Componets properties. Exactly attributes labels. I setup labels for each attribute, but no effect on rendering. Anybody have this problem?
I use JD 3.1<HR></BLOCKQUOTE>
By the way: setMaximumFieldHeight() don't work too!
null

JSTL Problem with attribute names

(Sorry for double posting. I've asked this question on javaranch, but did not get any answer. So I hope you can help me.)
The values I pass (request.setAttribute etc.) from my servlet to my JSP pages have often dots in their names like "my.variable" (Assume for the following that my.variable justs holds the String "hello").
I want to use <c:out > to print the value of "my.variable". Obviously the naming scheme collides with the expression language ("my" is not an object and certainly does not have the getter getVariable). Is there a way to work around this problem?

What? Why do you consider it a bad way? It's the only
naming scheme I know!
It's pretty straight forward to use the package name
to make your attribute names unique. See either Struts
or Java Petstore (org.apache.struts.Globals or
com.sun.j2ee.blueprints.waf.controller.web.util.WebKeys
as an example.Btw those Globals and WebKeys, are static finals or constants.
this is something completely different, and if you would really look good in stead of refusing some friendly advice you'll see that indeed the packages are separated with a . and the variables are with a _
so calm down and try to listen what we say.
If you give to java a x.y he'll thing y is a method of x or that y is a class in the package x
ok?
greetings

Native C++ metod, problem with attributes declared in other meth of Class

Hi guys, I have an interesting question for you.
I have a big C++ program, with many classes and methods..., and a java program which uses one of the methods(for the moment!) of a class.I compiled the java program, created a .h file, changed a little bit the C++ program with JNI functions...ok i stop talking, this is my C++ piece of program:
JNIEXPORT jobject JNICALL
Java_ProjectionType_setProjectionType(JNIEnv *env, jobject obj, jstring
projectionType,jstring mapState){
const char *str =(env)->GetStringUTFChars(mapState,0);
if (mapState == NoMap) {
reportError ("setProjectionType failed, no active map image");
return false;
if (mapState == MapDrawing) {
reportError ("setProjectionType failed, drawing already started");
return false;
env->ReleaseStringUTFChars(mapState,str);
// Check the projection type
const char *sti =env->GetStringUTFChars(projectionType,0);
if (projectionType < Orthographic ||
projectionType > PolarSouth) {
reportError ("setProjectionType failed, invalid projection type");
return false;
mapProjectionType = projectionType;
switch (mapProjectionType) {
case Orthographic:
viewportLeft = -180.0;
viewportRight = 180.0;
viewportTop = 90.0;
viewportBottom = -90.0;
break;
case PolarNorth:
viewportLeft = -90.0;
viewportRight = 90.0;
viewportTop = -90.0;
viewportBottom = 90.0;
break;
case PolarSouth:
viewportLeft = -90.0;
viewportRight = 90.0;
viewportTop = 90.0;
viewportBottom = -90.0;
break;
return true;
env->ReleaseStringUTFChars(projectionType,sti);
I think this is quite correct~. But here is my problem: all the attributes that I declared in other methods, such as NoMap, MapDrawing..., seem to be undeclared.
(The C++ program itself worked fine!)
Here is an example of the error messages I encountered:
`NoMap' undeclared (first use this function)
What do I have to do to make it declared?
Any answer willbe welcome
Philippe

You haven't given enough information for someone to see what is going wrong. However I suspect you may have a couple of problems.
1. You should realize that what you have written is a subroutine, not a method. It is not part of any object declaration.
2. You refer to "attributes I declared in other methods...". This sounds mixed up to me. What other methods of what?
3. Watch out. When you get to actually trying to run this from java, you will probably get linkage erros. I assume the code you posted is in a cpp file. You need to declare that you want C linkage; otherwise the compiler generates code that java can't understand.

Problem when attribute fails schema validation

I have got a problem during schema validation... let's say i have the following record:
<employee ssn="123456789D" recordNum="10">
<lastName>Bloggs</lastName>
<firstName>Joe</firstName>
<address>123 America Ave</address>
</employee>
apparently the attribute ssn will fail the schema, since ssn has to be all digit and length 9. Currently my schema vaildation process will keep on going to the next record if the current one fails, and write the current bad record to a error log file. the problem i have found is that if the ssn attribute validation fails, then there is no way for me to obtain the value of recordNum attribute (all elements can still be obtained).
just wondering if anyone had similar problem...
thanks

My understanding of schmemas was that once an error is found the whole document is considered invalid. Kind of like if the XML file format is bad.
If you don't mind me asking are you using DocumentBuilder to parse and is it giving you a SAXParseException?

OSB 11g - Conditional Branch Problem using attribute selection

Folks,
Came across a problem using OSB Conditional Branch using attribute selection, it fails & is most likely a bug. It was bug in ALSB2.5 but read in forums was fixed in ALSB 2.6
things I am doing -
1. In the Msg Flow, Conditional branch is @ the start of the flow
Xpath :- ./cm:processLineItem/cm:lineItem/@actionCode
In-Variable: - body
Label: - MODIFY
Operator:- =
Value:- 'Modify'
It fails with BEA-382000 error Code
<con:errorCode>BEA-382000</con:errorCode>
     <con:reason>
     com.bea.wli.sb.stages.StageException: {bea-err}TYPE003: Runtime Type Mismatch
     </con:reason>
What could be wrong with Xpath or is it a Bug ?
Thanks,
Abhijeet

Hi Abhijeet,
Your case is re-producible. It may be a bug.
BTW, I have noticed a strange behaviour -
1. If there is only one attribute then it works fine (<cus:lineItem actionCode="Modify")
2. If the attribute which you want to refer in selection, is in the last then also it works fine (<cus:lineItem priorityRanking="201" quantity="1"> actionCode="Modify")
So I will suggest you to raise a case with support and meanwhile as a workaround I would suggest you to use actionCode attribute in the last of lineItem node. So your XML should be like -
<soapenv:Body xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/">
<cus:processLineItem xmlns:com="http://vodafone.com.au/ebo/vha/CommonComponents" xmlns:cus="http://vodafone.com.au/cm/ocv/ebm/CustomerOrder" xmlns:cus1="http://vodafone.com.au/ebo/vha/CustomerManagement">
<com:header>
<com:createDateTime>2008-09-29T11:49:45</com:createDateTime>

<com:webUser>test</com:webUser>
<com:channel>string</com:channel>
<com:applicationName>string</com:applicationName>
<com:businessIdentifier>string</com:businessIdentifier>

<com:entityIdentifier>string</com:entityIdentifier>

<com:sourceIdentifier>string</com:sourceIdentifier>

<com:statusCode>Success</com:statusCode>

<com:messageIdentifier>string</com:messageIdentifier>
</com:header>
<cus:lineItem priorityRanking="201" quantity="1" actionCode="Modify">


<cus1:product ID="string">

<cus1:attribute>
<com:name>test</com:name>
<com:value>test1</com:value>
</cus1:attribute>
</cus1:product>
</cus:lineItem>
</cus:processLineItem>
</soapenv:Body>
Regards,
Anuj

Aggregate DB index problems in Attribute changerun

Dear All,
The attribute changerun in Process chain is hanging regularly and if I check the SM50 process overview
I am able to see the SQL statement " CREATE BITMAP INDEX "SAPR3"."/BIC/F200378~080"
ON "/BIC/F200378" ( "KEY_1003788" ) PCTFREE 10" which is running in loops for hours.
We tried the SAP Note 728922 but only few days it worked and remaining days till we deactivate the aggregates the Attribute changerun will be hanging.
We can't afford deactivating and refilling of the aggregates as we have 7years of data
kindly suggest me how to Avoid this.
Regards,
Ugendhar

Dear Ajeet ,
The issue is coming most of the times with two aggregates which have just 2117144 and 4022462 records respectively.
I will try creating the copy of the aggregates yearwise but this may hinder the performance of the query when we view for several years as it may have to split across many aggregates.
Dear Krishna,
The problem is not with the Cube indexes and facing the issue in ACR which is not related to cube but the aggregates of it.
Thanks for ur suggestions and please guide with more inputs if possible.
Regards,
Ugendhar

Strange problem - session attribute

Hi,
This one is really strange. I have a search field on my form. When I enter a value in the field(say 'test data') and then click 'Submit', the search is performed. To retain the above search value on the screen, I put it into a session attribute and then pick it up again from there.
The problem is, when the screen reloads after the search, only the first part of the string i.e 'test' is displayed and data disappears. Is it that when get values of session attributes in a JSP, only the value till an empty space is considered?
I would be grateful if someone could let me know if they have ever experienced the same of if there is any documentation stating this. I am totally lost!
Thanks in advance.

Greetings,
Could you please include some of your code. It seems like somewhere along the process you are either losing or not storing all the data properly.
Thanks,
-J

Problems adding attribute via OID Self-Service Console

I have added a new attribute via the Oracle Directory Manager tool, but when I map it to the Self-Service Console and select the Predefined List from the User Interface Type, and add the values to the list, only the last value will show up as many times as entries are in the list. For example, the list has the values Y N R, and when updating the value in the Console, the attributes list values are R R R.
I have not found anything on Metalink indicating a bug. Has anybody have any experience with this problem?
any help will be appreciated.
thanks

Please open a SR on this issue or open a bug.
regards,
--Olaf

RAID perfomance problem

Hi! I'm using 2xWestern digital 80GB ATA100 7.2k rpm 8mb on raid 0. I'm using a modified BIOS to set the strip size at 16kb.
The problem: perfomance.Sandra2003 scores 30000KB/s X( (my brother on a plain ata controller scores the same on a single 120GB unit). Persons who I've met get 42000 ;(
Also the NTFS cluster size is currently at 32kb. Before it was at 4 kb but the scores didn't change.
I'm using the downloaded drivers from live update.
What do you think I should do?
Can I change the strip size without loosing all the data on the drives? (I would have to delete the array and setup it again).

Quote
Originally posted by maesus
1. Try with different strip size [backup ur data first]
2. Use VIA Raid Performance Patch
3. use PCI Latency Patch
4. Ask that person who scored 42000 and copy his/her settings over. :D
Reinstalling windows and all the programs I have now would take an entire day
I've installed both VIA patches. She uses a 32kb strip size and 64kb cluster size (that is very agressive since i'd lost 2GB).
Do you know if there is a recent driver from promise (not from msi),and where do I download? (on the promise support site I don't know which model should I choose).

Problems with attribute types, need help!

Hi, everyone! I have a problem. In the first time I have an attribute Number in DB. I use this attribute in JDeveloper form. But now I need that this attribute be with type VARCHAR2(30). I changed DB, after that changed Entity object in JDeveloper to VARCHAR2(30), but now when I try to save record and wrote this attribute as cat&dog (VARCHAR2(30)), there are error: it is not a number.
WHY it is so? What else I need to change?
BEST REGARDS

Make sure that your change has propagated all the way to the corresponding ViewObject.
If it has not you can manually assign the attribute to the correct data type by editing the specific ViewObject attribute in the editor's Overview panel

Problem with attributes in custom component

Hi All,
I am creating a simple jsf PanelGrid custom component....in that component I am adding two new attributes....
when I use that tag in my jsf application it is rendering fine in the client side....but the parent class that is PanelGrid component's attributes are not inheriting to my custom panelgrid....
can any one tell me how can I use the attributes of the super class PanelGrid??
Thanks...

Have you configured the attributes in the tag library descriptor?

JDev 10.1.3.0.4 studio problem: table attribute not found in @JoinTable

Hi,
I tried to compile a EJB3 application in JDev 10.1.3.0.4 Studio version, but the compiler
complains about the table attribute of the @JoinTable annotation.
The following code is not accepted:
* Retrieve the list of clubs this player played for
* @return List of clubs this player played for
@ManyToMany(fetch=EAGER,targetEntity=nl.company.contractadmin.Club.class)
@JoinTable(table=@Table(name = "CONTRACTPERIOD")
joinColumns = {@JoinColumn(name = "PYR_ID",
referencedColumnName = "ID")}
inverseJoinColumns = { @JoinColumn(name = "CUB_ID",
referencedColumnName = "ID")
public Collection getClubs() {
return clubs;
public void setClubs(Collection<Club> clubs){
this.clubs=clubs;
This code compiles in the EA release of JDev 10.1.3..
Can anyone tell me if there is something changed in the implementation of EJB3??

On the latest build of JDeveloper, the backspace key mapping is still NOT WORKING. I've been trying to fix this since the early access release, but I failed.
Backspace key is already mapped by default, but still not working.
Did I missed something during the installation here? The installation documentation said that I should only extract the archive then edit the jdev.conf file to set the JDK to use.
I already pointed my installed JDK and JDeveloper's JDK but the backspace is still not working.
Help!

Perfomance problems changingTransaction Attribute

Similar Messages

Maybe you are looking for