BW Queries/data sources

Hello:
I agree on the concept about 1 universe for 1 BW query as mentioned in earlier response..
In that case wil we have seperate connections while creating the universe for each query ?
So 5 different connections for 5 Universe -> 5 BW queries?
Or
Can we use 1 BW connection for all the 5Universes for 5 BW queries
Appreciate your input

Hi,
You will need 1 connection per BW Query.
Regards
Alan

Similar Messages

InfoProviders versus Queries Data Source for BO SAP Integration

Hi All,
I would like to know if there is any discussion or suggestion on commonly used choice of data source for BO with SAP Integration. I do know there is a best practices suggestion from Ingo on Webi and the choice is to use queries so that we could leverage the functionalities of calculated key figures, exit variables and etc. But as I see from another website it claims that this will slow down the access of data using queries.
Thanks
David

Hi Ingo,
Is there any slide presentations from anyone that has that kind of information? We are in the midst of implementation right now and are selecting the method. I need to come out with pros and cons of the choices that we have in order for us to select the best one even though from my gut feeling I would still prefer the Queries.
Especially it will also make use of aggregates and also data flow size might be fewer due to variable that exists. Compared to using MultiProvider, it might transfer the whole data lot and if we were to create parameter there in Crystal or something else it will filter data only from there onwards.
Please correct me if I am wrong.
Thanks
David Yee

Excel-2007 cannot connect to Oracle ODBC data source, Control Panel can.


I cannot make an ODBC connection from Exce-2007 to Oracle work. I am an expert Excel and VBA user (since 1994) and I have frequently used Excel to access ODBC databases, including Oracle (I have done this with Excel-2003 both with worksheet queries and have written VBA ADO-connection routines). And even though in Excel-2007 a worksheet ODBC query is supposed to be easier to create than in previous version of Excel, my connection fails. Any suggestions and all help are welcome and much appreciated.
DETAILS


1) What is my system? I am using Excel-2007 on Windows Vista x64 and Oracle server v.11g on my computer (all this is on my computer, no network issues).
2) Why use Excel with Oracle at all? I use Excel-2007 to access Oracle rather than Access-2007 (or any other application like TOAD, etc.) because I do engineering calculations with the data stored in Oracle. These calculations are easier done in Excel (I suppose that one alternative to this could be to use some sql or Access to get the data from the database, then store it as plain vanilla CSV file, then open this file in Excel, then do the math (the math involves complex optimisation algorithms), then save the results as CSV, then use some sql or Access to put the data back into the database. Howwever this does not strike me as a quick or neat solution. And after all Excel has been designed to access ODBC databases, so why not use it?)
3) What do I do in Excel-2007 that won't work? I create an ODBC link to Oracle that does not work. In Excel-2007 this is straightforward:

<ul>
<li>define an ODBC connection (Data tab --> From other sources --> From data connection wizard);</li>
<li>define a query on the worksheet -- that's it, this is all!</li>
</ul>

I start with creating an ODBC connection:
a) I choose an ODBC data source type: ODBC DSN
b) Excel-2007 displays the list of the available ODBC data sources. I see in it my Oracle database name and I select it.
c) Excel-2007 displays the Data Link Properties:
- the "Provider" has a list of the OLE DB drivers with preselected "Microsoft OLE DB Provider for ODBC Drivers". I keep this default selection.
- the "Connection" tab has a connection string "DSN=<my database name>" which I keep, it also has fields for the user name and the password, which I fill with the correct credentials.
- Finally there is a button "Test Connection", which when I click produces the following error message:
"Test connection failed because of an error in initializing provider. 
Unespecified error"

4) Additional food for thought:
a) In the above walk-through the only data, which I type, are the user name and password, everything else is selected from lists offered by Excel-2007, hence any possibility of typos being the cause of the problem can safely be discarded.
b) I can test the ODBC driver in the Control Panel and it shows that it can connect to the Oracle database:
- in Control Panel --> Admin Tools --> Data Sources (ODBC) on the "User DSN" tab I can see the list of the available ODBC data sources (same list as in Excel-2007, point 3b above) with the name of my database in it;
- selecting the name of my database from the list of the sources and clicking "Configure" button opens a tab with Data Source Name (same as in Excel-2007), TNS Service Name and User ID. I enter <user name>/<password> and click "Test Connection" button. A message "Connection successful" appears (just for a test I enter incorrect user credentials and "Unable to connect" message appears)
BOTTOM LINE


The procedure for using an ODBC connection from Excel is very simple, in the past I have created and used such connections numerous times with Excel-2003 and earlier on Win-XP and earlier. But now on Excel-2007 and Vista-x64 I cannot make it work.
Also, testing an ODBC connection driver is really easy and simple to be done in the Control Panel. There testing the same ODBC connection, which fails in Excel-2007, results in success.
I am frustrated by the simplicity of the problem and yet the persistant error. I have lost now two full days in failed attempts to make the simple procedure work and in searching the internet for answers.
All help is highly appreciated <img class="emoticon" src="images/emoticons/happy.gif" border="0" alt="" width="16" height="16" />,
Plamen


Did you find the solution to your problem?
If not, I think I may know.
Excel 2007 is a 32-bit application.
When installing 32-bit applications in a 64-bit environment, the "default" location is:
C:\Program files (x86)\...
Excel then launches from this location.
However, when it connects to Oracle and passes the name of the calling program, Oracle attempts to "interpret" the value of (x86) as if
the value within the parenthesis are being passed as a reference. Of course, iOracle doesn't find anything, so the result is (),
and then it cannot return the connection info back to the calling application.
I corrected it by installing Excel in C:\Program Files\, and once launched from that location, it works the same as on the 32-bit machines.
However, at my location, they are FORCING Excel to be installed in the(x86) location.
What I'm trying to discover now is:
Is it possible to flag Oracle to NOT process the embedded variables?
Or, is it possible to assign a variable in Oracle such that x86 = "(x86)", so that the end result is viable?
Have you had any luck with your installation?
thanks,
Paul

How to create olap cube using Named Query Table in Data source View

I Create on OLAP Cube using Existing Tables Its Working Fine But When i Use Named Query Table with RelationShip To other Named query Table It Not Working .So give me some deep Clarification On Olap Cube for Better Understanding
Thanks

Hi Pawan,
What do you mean "It Not Working"？ As Kamath said, please post the detail error message, so that we can make further analysis.
In the Data Source View of a CUBE, we can define a named query. In a named query, you can specify an SQL expression to select rows and columns returned from one or more tables in one or more data sources. A named query is like any other table in a data source
view (DSV) with rows and relationships, except that the named query is based on an expression.
Reference:Define Named Queries in a Data Source View (Analysis Services)
Regards,
Charlie Liao
TechNet Community Support

11i EBS XML Publisher Report with Multiple Data Source

I need to create XML Publisher report in 11i EBS pulling data from another 10.7 EBS Instance as well as 11i EBS in single report.
I am not allowed to create extract or use db links.
My problem is how to create Data Source Connection using Java Concurrent Program.
The approach I am trying is
1. create Java concurrent program to establish connection to 10.7 instance.
2. Will write the SQL queries in Data Tempalete with 2 Data Source 1 for 11i EBS and 2 for 10.7 EBS
3. Template will show the data from both query in 1 report..
Is there any other way to proceed using datasource API...
thanks

option1:
The query should be same @ detail level, only the template has to be different for summary and details.
@runtime, user can choose to see the detail/summary
Disadvantage, if the data is huge,
advantage , only one report.
option2:
create two separate reports summary and details
and create diff data and diff layout and keep it as different report
Advantage, query will perform based on the user run report, summary/detail, so that you can write efficient query.
Dis advantage , two reports query/template to be maintained.

Diff b/w infosource and data source

hai all
can any one tell me abt the difference b/w infosource and datasource in b.w?.and also abt the infopackage and how the later relates with the former?.
thanxs
sri

Hi Srinivas,
Data source: It provides the data of source system,created in source systenm side and then replicated in BW side.Itis one which provides the data available in the tables to the BW System.A group of fields which are coming from one or more tables is called as datasource.Data will not be resided in the datasource.It will be present in the Tables.Using Datasource we extract the data into BW.
_TYPES OF DATASOURCE: _
datasource for transction data
datasource for char attributes
datasource for texts
datasource for hierarcy (att,texts, hier are master data datasources)
InfoSource: It is the source of meaningfull data for data target...created in BW side..it is basically logical pool of data ( no physical storage) from one or different data source... It is on the BW which contains all the InfoObjects. It contains transfer structure and communication structure. Transfer structure is actually the structure similar to the R/3 Datasource structure and comm. structure contains all the InfoObjects.
in 3.5 it has two types of infosource... flexible update(transction data) and direct update (master data)
core parts: communiation structure and transfer structure...
and infosourc in 7 you logicallcy group striaght infoobjects......
*PLS Search SDN for all ur Basic BW Queries*
*pls assign points,if info is useful*
Regards
CSM reddy

Bursting using a Concatenated Data Source

Greetings,
I am trying to burst a report using a concatenated data source. I have a bursting SQL query set up that works fine, but I am unsure how to handle the "Split By" option. I have a data model that pulls data from multiple queries, like this:
DETAILS_A
select * from table_a
where payee_id in (:p_payee_id)
DETAILS_B
select * from table_b
where payee_id in (:p_payee_id)
So, if I choose to "Split By" Details_A_Row/Payee_ID, then the data from Details_A is split appropriately, but the reports I generate don't split the information from Details_B properly. I end up with reports where one payee has information for another payee from the Details_B data source. How can I specify that the report should also split/filter the info from Details_B?
To put it another way, what I'd really like to do is go through a list of parameters (payee IDs in this example) and generate a report for each parameter. Is Bursting the most effective way to do this?
From searching the forums, it seems like I might be able to accomplish this using Data Templates as my Data Model instead of SQL Queries. Am I on the right track with that? If so, you guys have any helpful links on how to create Data Templates?
I am using BI Publisher version 10.1.3.4
Any help is appreciated!
Martin

For data template samples, check the following out and then download the zip file (you may not be able to run the reports, but you can view the code)
http://blogs.oracle.com/xmlpublisher/2009/06/data_template_progression.html
Thanks,
BIPuser

Any examples of a data template using multiple data sources?

I'm looking for an example report using multiple data sources. I've seen one where they do a master/detail but I'm just looking to combine results in sorted order (sorted across all data sources). The master/detail used a bind variable to link the two defined queries, I'm thinking what I want won't have that, so I'm lost on how to make that happen. I have reports using multiple sql queries and there is a way in the data source pulldown to tell it to combine the data sources. It appears to be a more manual process with data templates, if it's even possible.
Any pointers/links would be appreciated.
Gaff

Hi Vetsrini :
That's just it. Mine is simpler than that. There is no master/detail relationship between the two queries. I have the same exact query that I run in two databases and I want to merge the results (ordered, by, say eventTime) in one report. So I think my results are going to be two separate groups (one for each data source) which I'll have to let BI merge vis XSLT or whatever it uses. That's fine for small result sets but for larger ones, it would be nice if the database did the sorting/merging.
Gaff

Report based on custom Query - Question about Data Source

Hello,
I am trying to create Crystal reports (EBO R3.1, Oracle 11g Database, CR2008) based on custom queries. Tried using different kinds of Data Sources but getting stuck once I export them to Infoview.
1. Using Oracle Server: This method works fine on my local & Infoview. But since I have lot of reports, I do not want the DB username/password associated to the report. I neither want BO to prompt username/pwd to my user.
2. Using ODBC(RDO):
Again, works fine on my local. Once I export to Infoview, it fails with this error: "Failed to open the connection"
Which makes sense, because the DSN is created on my local machine and not on the server.
How and what do I create on the server, so that the reports run?
I do not know if using ODBC(RDO) is even the right way for creating custom SQL reports.
I have been reading about Business Views, but not clear on how good they work.
My goal is to create multiple reports, using custom SQL queries which use some kind of connection to the database that can be managed on the
server in one place, and used by multiple reports.
Can any of you please help me here
Thanks,
Sowmya.
Edited by: SBat22 on Apr 13, 2011 5:58 PM

Thanks for moving it to the right forum!!
But, if I setup in CMC, a default username/password that the report needs....what would happen if my password changes?
I would need to go to each and every report and update the password.
I am looking for using one common "connection" for multiple reports. So that, when passwords change, all I update is in one central location.
Can anyone please help!
Thanks,
Sowmya.

Trying to change the data source for a Crystal Report.

The method below represents my best attempt to programatically change the data source of a Crystal Report. The goal is to have a routine that will update the data source for reports after they have been distributed to production servers. So far I have not been successful in saving the report back to the CMS. No exceptions are thrown, but when I view the Database Configuration of the report in the CMC nothing has changed.


Am I missing a step, or is there another way to accomplish this?


Thank you.

<hr />
<pre>
private void test(String reportName)
 throws SDKException, ReportSDKException, java.io.IOException
 IInfoObjects newInfoObjects;
 IInfoObject reportObj;
 ReportClientDocument clientDoc = new ReportClientDocument();
 DatabaseController dc;
 PropertyBag pBag;
 PropertyBag logonProps;
 ConnectionInfo newConInfo;
 ConnectionInfo oldConInfo;
 ConnectionInfos conInfos;
 int connOptions = DBOptions._ignoreCurrentTableQualifiers + DBOptions._doNotVerifyDB; //0;
 Fields connFields = null;
 String queryStr = "Select * From CI_INFOOBJECTS " +
 "Where SI_NAME='wfr.rpt' AND SI_KIND='CrystalReport' AND SI_INSTANCE=0";
 newInfoObjects = getCms().executeQuery(queryStr);
 if(newInfoObjects.size() > 0)
 reportObj = (IInfoObject)newInfoObjects.get(0);
 try
 clientDoc = getCms().getReportAppFactory().openDocument(
 reportObj
 , OpenReportOptions._refreshRepositoryObjects
 , java.util.Locale.US);
 dc = clientDoc.getDatabaseController();
 conInfos = dc.getConnectionInfos(null);
 for(int i = 0; i < conInfos.size(); ++i)
 oldConInfo = (ConnectionInfo)conInfos.getConnectionInfo(i);
 newConInfo = (ConnectionInfo)oldConInfo.clone(true);
 pBag = newConInfo.getAttributes();
 pBag.putStringValue("QE_ServerDescription", "alio");
 logonProps = new PropertyBag();
 logonProps.putStringValue("Trusted_Connection", "false");
 logonProps.putStringValue("Server", "alio");
 pBag.put("QE_LogonProperties", logonProps);
 newConInfo.setUserName("admin");
 newConInfo.setPassword("password");
 dc.replaceConnection(
 oldConInfo
 , newConInfo
 , connFields
 , connOptions);
 catch(ReportSDKServerException Ex)
 String msg = "A server error occured while processing the " + reportObj.getKind()
 + " object, " + reportObj.getTitle() + " (" + reportObj.getCUID() + "), from the CMS.";
 Utility.errorOut(msg, Ex);
 catch(Exception Ex)
 String msg = "An error occured while processing the " + reportObj.getKind()
 + " object, " + reportObj.getTitle() + " (" + reportObj.getCUID() + "), from the CMS.";
 Utility.errorOut(msg, Ex);
 finally
 clientDoc.save();
 getCms().commitToInfoStore(newInfoObjects);
 clientDoc.close();
</pre>
Edited by: Mark Young on Sep 10, 2009 2:16 PM

<style type="text/css">
/<![CDATA[/
 body
 font-size: 1.125em;
 font-family: helvetica,arial,"sans-serif";
 .code{font-family: "courier new",courier,mono,monospace}
 .bi{font-style: italic; font-weight: bold;}
/]]>/
</style>
Justin,

Thank you for the reply. Time constraints have not allowed me to post back to this tread
till now. I will try your suggestion. My assumption is that Save the report back to the
info store refers to IInfoStore.commit(IInfoObjects).


I'm afraid that I do not understand why I don't want to change the report client document,
or why successfully exporting the report with the new login/password is not what I
want to do. Any explanation on that statement would be appreciated.


I did find a way to accomplish my goal. It involved adding the SSOKEY property to the
logon property bag. Below you'll see my revised code which modifies the report logon and
server. I have no idea what
this does, and SAP support has not been able to tell me why it works. However, what I
discovered is that if I changed the report option, Database Configuration -> When
viewing report:, in the CMS to Use same database logon as when report
is run from Prompt the user for database logon, then the
SSOKEY property had been added to the logon property bag having an empty string as its
value. This allowed me to successfullyupdate and save the modified logon back to the CMS.


So I took a chance and added code to always add the SSOKEY property with an empty string
as its value, and I could then successfully modify and save the report's logon info
and server. Again, I don't know what this means, but it has worked so far. If anyone has
some insight or comments, either are welcome. Thank you in advance.

 
<hr />
<pre>
private void changeDataSourceOfAWFCrystalReports()
 throws SDKException, ReportSDKException, java.io.IOException
 IInfoObjects newInfoObjects = null;
 IInfoObject reportObj = null;
 IReport curReport = null;
 ReportClientDocument clientDoc = new ReportClientDocument();
 DatabaseController dbController;
 PropertyBag pBag;
 PropertyBag logonProps;
 ConnectionInfo newConInfo;
 ConnectionInfo oldConInfo;
 ConnectionInfos conInfos;
 int connOptions = DBOptions._ignoreCurrentTableQualifiers + DBOptions._doNotVerifyDB;
 Fields connFields = null;
 String outputStr;
 int numOfReports;
 int numOfQueryPages;
 double progressIncrementPerPage = 30;
 int progressIncrementPerReport = 0;
 // Path query to reports is in a .properties file.
 String queryStr = getAppSettingsFile().getWscAwfCrystalReportPathQuery();
 try
 // Executes IInfoStore.getPageingQuery() and generates a list of queries.
 getCms().setPathQueryQueries(queryStr, 100);
 numOfQueryPages = 0;
 // Gets a List<String> of the IPageResult returned from IInfoStore.getPageingQuery().
 if(getCms().getPathQueryQueries() != null)
 numOfQueryPages = getCms().getPathQueryQueries().size();
 if(numOfQueryPages > 0)
 // Use 30% of progress bar for the following loop.
 progressIncrementPerPage = Math.floor(30.0/(double)numOfQueryPages);
 for(int queryPageIndex = 0; queryPageIndex < numOfQueryPages; ++queryPageIndex)
 // Gets the IInfoObjects returned from the current page query
 newInfoObjects = getCms().getPathQueryResultSetPage(queryPageIndex);
 numOfReports = newInfoObjects.size();
 if(newInfoObjects != null && numOfReports > 0)
 progressIncrementPerReport =
 Math.round((float)Math.floor(progressIncrementPerPage/(double)numOfReports));
 for(int reportIndex = 0; reportIndex < numOfReports; ++reportIndex)
 reportObj = (IInfoObject)newInfoObjects.get(reportIndex);
 curReport = (IReport)reportObj;
 clientDoc = getCms().getReportAppFactory().openDocument(
 reportObj
 , OpenReportOptions._refreshRepositoryObjects
 , java.util.Locale.US);
 dbController = clientDoc.getDatabaseController();
 conInfos = dbController.getConnectionInfos(null);
 for(int conInfosIndex = 0; conInfosIndex < conInfos.size(); ++conInfosIndex)
 oldConInfo = (ConnectionInfo)conInfos.getConnectionInfo(conInfosIndex);
 newConInfo = (ConnectionInfo)oldConInfo.clone(true);
 pBag = newConInfo.getAttributes();
 pBag.putStringValue(
 "QE_ServerDescription"
 ,getConfigFile().getDBDataSourceConnections());
 logonProps = new PropertyBag();
 logonProps.putStringValue("Trusted_Connection", "false");
 logonProps.putStringValue("SSOKEY", "");
 logonProps.putStringValue(
 "Server"
 ,getConfigFile().getDBDataSourceConnections());
 pBag.put("QE_LogonProperties", logonProps);
 newConInfo.setUserName(getConfigFile().getUNVConnectionUserName());
 newConInfo.setPassword(getConfigFile().getUNVConnectionPasswordDecrypted());
 dbController.replaceConnection(
 oldConInfo
 , newConInfo
 , connFields
 , connOptions);
 newConInfo = (ConnectionInfo)conInfos.getConnectionInfo(conInfosIndex);
 } // end for on conInfosIndex
 clientDoc.save();
 } // end for on reportIndex
 } // end if on newInfoObjects
 } // end for on queryPageIndex
 } // end try
 catch(ReportSDKServerException Ex)
 // handle...
 catch(Exception Ex)
 // handle...
 finally
 getCms().commitToInfoStore(newInfoObjects);
 if(clientDoc != null)
 clientDoc.close();
</pre>

Data Source Missing in Publication

HI experts,
I've got a business objects publication which uses 1 crystal report as the source data and 1 crystal report as the dynamic recipients. When run seperately both these crystal reports work fine but when I try to schedule the publication I keep getting the error message 'Unable to connect. Incorrect logon parameters'.
The source data crystal report uses 2 different data sources and when I look at the database logon parameter within the report I can see them both and they have the correct passwords. When I look at the database logon parameter of the source data within the publication I can only see 1 datasource.
I'm assuming my problem is something to do with the missing datasource but I don't know how to view it within the publication.
What am I doing wrong?
Thanks
Stuart.

Hi Mansi and Rathi
Will creating a new infopackage or replicating hamper anything that is already existing. I am new to BI and actually was asked only to work in queries. I dont want to take a chance, just in case.
I have 3 infopackages already created for this data source.
Thanks and Regards
Madhu

Error During Replication of Data Source in BI 7.0 SAPKW70017

Hi Experts
We have recently imported Support Pack SAPKW70017 into our quality
system. The Quality System Client setting is
- No changes Allowed
- No change to Repository and cross-client Customizing objects
After upgrade when ever we are trying to replicate data sources we are
getting the error. "Changes to Repository or cross-client Customization
not permitted".
This error was certainly not there before updating to SAPKW70017.
Interestingly if we change client setting to
- Automatic Recording of Changes.
- Changes to the Repository and cross-client Customizing permitted
and try to replicate data source. System does not prompt for any change
request.
To overcome this, we tried to import the 'data source in BW' via change
request from development. The replication completes successfully but
time stamp does not matches with source. And we end up in
error "DataSource 0FI_AR_4 has to be replicated (time stamp, see long
text)" in scheduler.
Please advise.
Thanks & Regards
Sachchida Nand Dubey

hi,
I am looking forward to implement SAPKW70017, I need to know the CONS of SP17, what was the reason which made you implement SP17, just to have updated Version of SP or did you face with any problems and decided to update the SP? what was the previous SP which you had,
We have problems in F4 search help when running queries, SAP asked us to implement latest which is SP17, but wanted to make sure what would be the CONS before we do it, we are having SP15 in BI 7.0
Please update the post for this.

Data Source created by Infoset based in two joined tables

Hi Forum,
I want to do a segmentation in the Segment Builder based in a attribute list.
The tablas i join are: BUT000 and ADRC. Everything is ok when the infoset is created and generated.
Well, the next step i do is create the DATA SOURCE. When i edit source data i fill these fields:
Origin type --> 03 Infoset
Name of Infoset --> ZINFOSET
BP --> BUT000-PARTNER
FM --> ZCONVERSION_MKT (this function convert the guid of my partners in a marketing guid, it works because the segmentation based in a infoset with a table works)
Sampling Infoset --> ZINFOSET
Sample --> BUT000-PARTNER
Object --> BUT000-PARTNER_GUID ¿?¿?¿?¿?¿
Partner Function --> Empty
Description --> my description
So, when i save this data...appears an error with this description:
Techn. Name (BUT000) of InfoSets "ZDEF2" is not transparent table
with this diagnosis:
Each InfoSet in the IObject scenario must have a unique transparent retraction table assigned to it
I need a object field, but i dont know what field choose...in the Segment Builder i choose the list and i create a few filter to try some queries but i doesnt work.
Any suggestion or idea¿? or example?
Regards,
Mon
Edited by:
Mon

Hi!
It's solved. The mistake was on the connection of the tables:
BUT000 - BUT020 - ADRC
Thanks and regards,
Mon

Data Sources , Data Providers to a Single WebI Report

Hi Guru's ,
I am a newbie..!! I am going through the documents related to a creation of WebI Report and got to know that we can combine two or more data sources (or) two data providers(Queries) in creation of a single WebI Report . However,I am still confused with this incoming stuff for creating a WebI Report .
Could some one explain me what type of combinations can give as an input for creating a WebI report ?
For example, one can link two or more queries and create a single report by using merge dimension concept !! If so, Will WebI accept data coming from 2 different universes, data coming from 2 or more different data sources like SQL data and SAP ecc ?
Please explain this as I am new to this !!

Hi
Common in the sense relevant data in the two objects with same data type.
See chapter 24 in the attached pdf more on merge dimension and how it will work in webi.
http://help.sap.com/businessobject/product_guides/boexir4/en/xi4sp4_ia_desktop_en.pdf

ODBC Data Source Won't Recover from Connection Errors

I have several ODBC data sources to a mainframe and want to
use the "Maintain Connections" setting as it dramatically improves
performance under heavy load. Problem is that we frequently have
path problems between the CF server and the mainframe 180 miles
away. With the Maintain Connections checked, the data source
doesn't consistently start working again after the path problem
clears up. The path problem usually only lasts a minute or two, but
SOME queries through the affected data sources continue to raise
errors until the ODBC Service is cycled. The SOME aspect is really
weird because not every hit through that data source reports an
error after the path comes back up. I've had side-by-side users hit
the same page using that DS and one consistently get's an error and
the other works fine. Restarting the ODBC service clears the
problem so it works fine for everybody. Of course that breaks all
the connections that were still working so I just affected a bunch
more users. Silver lining there is that at least nobody can claim
preferential treatment right. It's also kind of bizzare that I can
use the CF Administrator to verify all data sources and they
usually all report that they are working fine even while some
connections are still reporting problems.
The specific ODBC drivers we are using are Neon Shadow
Drivers (Multi Threaded). I'm thinking about using the Admin API to
hit the Disable Connections and then re-enable it.
Any constructive, tactful input for addressing this problem
would be appreciated.

spikehenning wrote:
> I have several ODBC data sources to a mainframe and want
to use the "Maintain
> Connections" setting as it dramatically improves
performance under heavy load.
> Problem is that we frequently have path problems between
the CF server and the
> mainframe 180 miles away. With the Maintain Connections
checked, the data
> source doesn't consistently start working again after
the path problem clears
> up. The path problem usually only lasts a minute or two,
but SOME queries
> through the affected data sources continue to raise
errors until the ODBC
> Service is cycled. The SOME aspect is really weird
because not every hit
> through that data source reports an error after the path
comes back up. I've
> had side-by-side users hit the same page using that DS
and one consistently
> get's an error and the other works fine. Restarting the
ODBC service clears the
> problem so it works fine for everybody. Of course that
breaks all the
> connections that were still working so I just affected a
bunch more users.
> Silver lining there is that at least nobody can claim
preferential treatment
> right. It's also kind of bizzare that I can use the CF
Administrator to verify
> all data sources and they usually all report that they
are working fine even
> while some connections are still reporting problems.
>
> The specific ODBC drivers we are using are Neon Shadow
Drivers (Multi
> Threaded). I'm thinking about using the Admin API to hit
the Disable
> Connections and then re-enable it.
Do you have a JDBC Type 4 alternative for these drivers? I
got rid of
all my datasource problems long ago when I got rid of ODBC
altogether.
Jochem
Jochem van Dieten
Adobe Community Expert for ColdFusion

BW Queries/data sources

Similar Messages

Maybe you are looking for