Contents
Did You Know?
Tips
Tip One: How to Create Watcher and Approver in the Approval Flow for Specific Catalog Items
Tip Two: Troubleshooting Data Loads between Ariba Analysis and Ariba Contract Management/Ariba Contract Workbench
Archive
Feature Article: Ariba Buyer 8.2.2: Detailed Steps to Purge and Archive
Featured Resource: Fixed Defect "Requires Action" Flag
Top Tech Notes
Top Hot Issues
Top FAQs
Ratings and Suggestions


Did You Know?
Remedy Error Encountered When Using the "-norestart" Command-Line Option for Startbuyer
By Fredrik Winnerhed, Senior Technical Support Engineer

Ariba Buyer 9r1: Using the command line option, -norestart for startbuyer, without the NodeManager running, for local development commonly makes restarts faster. However, it can produce the following error message:


Tue Jan 20 09:52:02 PST 2009 (T2) (util:WARN) [ID2935]: Server shutting down with exit code: 2.  Shutdown message: Unable to initialize node manager: ariba.util.nodemanager
.NodeManagerInitializationException: Unable to initialize nodemanager Cluster

If this error occurs, correct the problem by setting the NodeManagers.Cluster.DisableMultiNode parameter in the parameters.table to "true" instead of the default value, "false".


Service clients Ariba
Access https://connect.ariba.com to connect!

ASP Award Logo


Tips

Tip One
How to Create Watcher and Approver in the Approval Flow for Specific Catalog Items
By Sailaja Lingamaneni, Consultant Product Solutions Engineer

Ariba Buyer 8.2.2:  While developing a catalog, you can create a watcher and approver (different users) in the approval flow. Add the watcher and approver, with specific roles, while creating a requisition and selecting the items from the specific catalog.  

Because you cannot specify the same commodity code twice within the Server\config\variants\var1\data\CommodityApprovers.csv file, use two Comma Separated Value (CSV) files:

  • The first rule uses the approver CSV file to add the approver to the approval graph, as soon as you select the particular commodity code.
  • The second rule uses the watcher CSV file to add the watcher to the approval graph, as soon as you select the particular commodity code.

Create those rules in Administrator-->Rule Editor-->Edit JavaScript Requisition Rules, adding each as a simple rule and designating it as the "Commodity CSV rule" type.

While creating the Commodity CSV rule, you can specify the CSV file and save it. When you attempt to edit this rule, if the software does not find that CSV file, it asks whether it can create a new CSV file, giving you the opportunity to do so.




Tip Two
Troubleshooting Data Loads between Ariba Analysis and Ariba Contract Management/Ariba Contract Workbench
By Kevin Dvorchak, Support Programs, Project Manager

Ariba AnalysisTM and Ariba Contract ManagementTM (ACM)/Ariba Contract WorkbenchTM (ACW):

 

In most scenararios, understanding why data is not pulled into Ariba Analysis or why data does not display the expected value(s) requires the same method of troubleshooting; start small and work backward.

 

All data loads, weather default configuration or custom loads, run by executing an Ariba Query Language (AQL) query against the source system (ACM). Either pull this query from the Ariba Analysis Administrator Console in ASM4, or extract the query from a load definition file, such as the acmloads.xml file. Understanding which load pulls the data in question is essential; but, once established, the process of troubleshooting is usually the same.

 

After obtaining the AQL query that performs the load, place it in a text editor and break it apart so that each SELECT, JOIN, and WHERE condition appears on its own line:

 

Select

            A,

            B,

            C,

 

From Workspace as Project

Join Users, etc.

 

 Where

 

A=x and

B=y and

C=z

 

Then, from ACM inspector, start with as basic a query as possible and attempt to pull the record that contains the value(s) in question, often:

 

Select InternalId from Workspace where InternalId='123456'

 

This ensures an active record that can be returned. Next, take the load query that you broke apart and add an AND condition to make the query unique to a single record that is in question. 

 

Select a,b,c from Project where a=x etc... and  Project.InternalId='123456'

 

Essentially, this executes a data load, but one that you make specific to a single record.  In almost all cases, you need not actually run a data load to see what gets pulled. Simply, run the data loads AQL query from ACM. If the record returns, you know it will get pulled into Ariba Analysis. If the question regards the data returned, you can view how it gets returned in the query, and manually drill into the objects to better understand exactly what gets returned.  When the data does not get returned, start breaking the AQL query you formatted, removing lines one-by-one.

 

  1. Start by removing all but a single object from the select clause. If your select statement calls for the following, limit the statement to InternalId:

    Select InternalId,ProjectId,Owner,NextVersion, etc...

  2. Remove the conditions from the where clause, with the exception of the condition that limits it to your single record, for example:  "and Project
    .InternalId='123456'
    "

As an example, if the AQL query is formatted: 

select Project.InternalId from Workspace where x=y and a=c and c=r and Project.InternalId='123456'

  1. Run selectProject.InternalId from Workspace where x=y and a=c and Project
    .InternalId='123456'
    (Removing "and c=r")
  2. Run select Project.InternalId from Workspace where x=y and Project
    .InternalId='123456' (
    Removing "and c=r" and "and a=c")
  3. Run select Project.InternalId from Workspace where Project.InternalId='123456' (Removing "x=y", "and a=c", "and c=r" and "and"), which is your original query.  

If you remove all conditions in the where clause and the record still does not get returned, start removing any joins one-by-one as you removed the conditions for the where clause. By following this basic troubleshooting process, you can gain a good understanding of the method by which the data gets pulled.




Archive
Access the Ariba Tech Connect Archive


 
Volume 10 - August 2009
Feature Article
Ariba Buyer 8.2.2: Detailed Steps to Purge and Archive
By Bhaskar Bodahanapati, Team Captain, Technical Support

Use these easy-to-follow steps to complete the Archive and Purge process in Ariba BuyerTM 8.2.2 effectively. This long process usually takes a few months. 

  1. Backup the Ariba Buyer database (DB) of the instance in which to run these scripts.
  2. Get the PRE-PURGE numbers from Inspector.

    1. This important step depends on your requirement. Define the objects to purge in an Extensible Markup Language (XML) file. (See the Ariba Buyer Configuration Guide Version 8.2 for a detailed description of that file and its purpose.) At this point, you can already catch quite a few issues. They usually appear on the WHERE conditions. Make sure to get the number of objects corresponding to exactly the same query from the XML file:

      SELECT COUNT(*)
      FROM ariba.purchasing.core.Requisition
      WHERE (StatusString != 'Canceled'
      AND StatusString != 'Composing')

    2. Verify that all of these objects are in PurgeState = 0:

      SELECT COUNT(*)
      FROM ariba.purchasing.core.Requisition
      WHERE (StatusString != 'Canceled'
      AND StatusString != 'Composing')
      AND PurgeState = 0

    3. Verify that the PURGE_CANDIDATES table is empty!

      1. Go to inspector > Purge Browser: nothing should display
      2. Go to inspector > Query Statement and run the following query, which should return a message similar to "Cannot find class PURGE_CANDIDATES":

        SELECT COUNT(*) FROM PURGE_CANDIDATES

  3. Run the PREVIEW phase as indicated in the previously mentioned documenation.
  4. Verify the candidate objects by starting Ariba Buyer and going to inspector/PurgeBrowser. Here, you frequently see inconsistencies between the numbers from the purge browser and the queries you ran before the PREVIEW phase. This can happen because the PREVIEW phase numbers also include dependencies between objects that your queries could not see, necessarily. At this point, the DB is not changed, and no object is marked. Also, get some numbers from the PURGE_CANDIDATES table, and verify that the PURGE_RECORD (which will contain the references to all objects we want to purge and archive) is empty. 

    + SELECT COUNT(*)
    FROM PURGE_CANDIDATES

    + SELECT COUNT(*)
    FROM PURGE_RECORD

    COUNT(*)
    0

  5. Run the MARK phase as described in the previously mentioned documentation.
  6. After the MARK phase completes:

    1. Verify the numbers, again, for each type of object as in Step 2. To do so, add the following condition to your queries:

      et PurgeState > 0

      At this point, Ariba Buyer cannot access any of the objects marked to purge. 

      SELECT COUNT(*)
      FROM ariba.purchasing.core.Requisition
      WHERE (StatusString != 'Canceled'
      AND StatusString != 'Composing'
      AND PurgeState > 0)

    2. Check the number of objects in the PURGE_RECORD and PURGE_CANDIDATES:

      + SELECT COUNT(*)
      FROM PURGE_RECORD

      + SELECT COUNT(*)
      FROM PURGE_CANDIDATES

  7. Run the ObjectArchive task as described in the previously mentioned documentation.

    Important: Once you start the ARCHIVE phase, you cannot run any part of the PURGE phase on the same DB! If you must run the PURGE again, you must start from a fresh DB backup.

    Usually, changing two things gets the objects purged and archived as fast as possible. They involve letting the task go as far as possible and reducing the number of times it must run. Use these at your discretion depending on your specific configuration and requirements.

    Set MaximumAllowedErrors = 10000;
    And set IgnoreBadFieldValue = true; 


    ObjectArchive = {
            Directory = objectArchive.10.08;
            IgnoreBadFieldValue = true;
            MaximumAllowedErrors = 10000;
            PurgeScope = Procurement;
            RunningTimeInMinute = 480;
            ScheduledTaskClassName = ariba.encoder.soap
               
    .base.ArchiveTask;
            SleepIntervalInMillis = 5000;
        };

  8. After the Archive task completes, check the logs and the number of entries in "failed" state by running the following query:

    select count(*) from PURGE_RECORD
    where STATE = 2


    Note:  At this point, no table entries should have "STATE = 2". If any appear, the Archive task probably is not finished yet.

  9. Run the FailedObjectArchive task. This task attempts to archive the objects in the error state, only. The following errors usually appear in the logs:

    1. Null Pointer Exceptions: usually indicate one of the following:

      • may require a code fix from Engineers
      • the existence of references to non-existing objects caused by a deleted user, deleted requisition, or a missing attachment

    2. Catalog Errors
    3. Wrong Binary (BIN) attachments in Archive directory: most of the issues occur on Microsoft Office Worddocument files (DOC) or Microsoft Office Excel Spreadsheet (XLS) files, again, depending on the specific configuration. In one case, the following fixed the problem (after finding that the problem occurred on DOC and XLS files):

      1. Get the numbers of documents to update:

        SELECT count(*)
        FROM AttachmentTab include inactive
        WHERE AT_CONTENTTYPE = 'application/octet-stream'
        AND AT_EXTENSION = 'doc'

        SELECT count(*)
        FROM AttachmentTab include inactive
        WHERE AT_CONTENTTYPE = 'application/octet-stream'
        AND AT_EXTENSION = 'xls'

      2. Update these objects, making sure that the content_type is supported by our code:

        UPDATE AttachmentTab
        SET AT_CONTENTTYPE = 'application/msword'
        WHERE AT_CONTENTTYPE = 'application/octet-stream'
        AND AT_EXTENSION = 'doc'

        UPDATE AttachmentTab
        SET AT_CONTENTTYPE = 'application/vnd.ms-excel'
        WHERE AT_CONTENTTYPE = 'application/octet-stream'
        AND AT_EXTENSION = 'xls'

      3. For real BIN attachments: some extensions are not supported, such as Microsoft Office Outlook Message (MSG) Format, "DOC=", or no extension. Usually, they represent a very small number.

      4. Zero kilobyte (0kb) attachments: some are 'real' 0kb files, some appear to be 0kb files because the attachment is missing from the customer file system. The following query identifies the 'real' 0kb files.

        SELECT Comments.Attachments
                .ContentLength,
            Comments.Attachments
                .Filename,
            Comments.Attachments
                .Extension,
            Comments.Attachments
                .ContentType,
            Comments.Attachments
                .StoredFilename
        FROM Requisition
        WHERE Comments.Attachment
            .ContentLength = 0

  10. Check the number of objects again to determine if any remain in the DB. A few objects may not be archived, but you usually find that acceptable, or can explain it. Also, after the Purge and Archive process, no entries should appear in the PURGE_RECORD table. 


    SELECT COUNT(*)
    FROM PURGE_RECORD

    COUNT(*)
     
 
 

Featured Resource
Fixed Defect "Requires Action" Flag
By Derek Matthews, Lead Knowledge Architect

The Fixed Defects portlets on the connect.ariba.com Product Info pages have a new column entitled "Requires Action." This new column highlights fixed defects that require action from an administrator to take advantage of the fix or new feature. Click the column header to sort these fixed defects together. Click the View All Fixed Defects that Require Action in One Document link to see the full details on one page.



 
 
Top Tech NotesLa suite 
Questions about Safe Customization in 9r1Ariba Buyer
Error Occurs while Running the IVT Tool on 9r1 UpstreamAriba Category Management
Start of ASM Server Fails With "Error loading metadata: java.lang.NullPointerException"Ariba Analysis
After Applying SP32 to the Server/ariba/oracle/R12 Folder, Subsequent SQL Scripts Are Not CreatedAriba Buyer
Users Must Enter an Additional ID and Password to Log into InspectorAriba Buyer
 
Top Hot IssuesLa suite 
Attachments Are Not Copied During the Data Migration to 9r1Ariba Analysis, Ariba Buyer, Ariba Contract Management, Ariba Sourcing
Microsoft Performance DefectAriba Analysis, Ariba Buyer, Ariba Category Management, Ariba Sourcing
Error in the Log File Saying That ariba.expense.server.AttendeeMigrator Cannot Be Found, after Applying 9R1 SP5Ariba Buyer
Cannot Submit Approvables after Installing 9r1 SP5Ariba Buyer
Since the Release of AN45, PCard Numbers Are Not Sent in Email Orders. Why?Ariba Supplier Network
 
Top FAQsLa suite 
What Log Categories Can I Enable to Monitor Performance? How Do I Interpret the Output?Ariba Buyer, Ariba Category Management
How Can I Measure and Analyze the Response Times of AQL Queries?Ariba Buyer, Ariba Category Management
Is Ariba Discovery the Same as the Public RFX? Ariba Supplier Network
What Are the Benefits of Ariba Advantage? Ariba Supplier Network
What is Ariba's policy for resolving issues found in Ariba Discovery?Ariba Supplier Network
 
 


Ratings and Suggestions
Ariba Customer Support welcomes your comments and participation in Ariba Tech Connect. Rate this issue, make content-related comments, or suggest a topic for a future edition. Thank you for your time and willingness to help us center the content on the information you need and want.
What is your satisfaction level with this issue of the Ariba Tech Connect bulletin?
<--Very Dissatisfied
Very Satisfied-->
12345678910

Please provide us with any additional comments and suggestions.
     

     Problems Submitting this Form?     

Designated Support Contacts for Test Company (Ariba)La suite 
PrénomNomFonctionE-mailTéléphone
TESTCUSTOMER1matthewsderek@gmail.com1234566789

Service Request 365 Day Performance Metrics SummaryLa suite 
Team # of SRs Created # of Closed SRs # of SRs Closed <1 Day # of SRs Closed <2 Days # of SRs Closed <7 Days Average Resolution Time (Hours) # of SRs Closed on Initial Contact
Application Maintenance 0000000
End User Support 31 3127273051.166
Post Deployment Maintenance 0000000
Premium Support 1 11111.570
Professional Services 0000000
Support technique 5 55555.631
Total3737333336 7
Percentage89.19%89.19%97.30%18.92%

Customer Satisfaction 365 Day Metrics SummaryLa suite 
QuestionAverage
For this service request, how satisfied were you with Ariba Customer Support?n/a
Initial speed of answer/response was timely.n/a
Support representative demonstrated product knowledge competence.n/a
Support representative communicated status of my issue at each stage.n/a
Support representative provided a resolution/workaround in a timely manner.n/a
Support representative was courteous and professional.n/a
Resolution/workaround provided was relevant, effective, and easy to understand.n/a
 
Related MetricsResult
Did the Ariba Customer Support representative provide exceptional service to you in resolving service request? (Number of Yes; No responses)0; 0
Number of Surveys <=6; Percentage of Surveys <=60; 0.0%
Number of Surveys >=8; Percentage of Surveys >=80; 0.0%
Number of Surveys0
Number of Invitations0


 


Manage subscriptions to this bulletin.

Politique de confidentialité

� 2009 Ariba Inc. All Rights Reserved

Microsoft, Microsoft Office, Microsoft Office Word, Excel and Microsoft Office Outlook are trademarks or registered trademarks of Microsoft Corporation or its subsidiaries in the United States and other countries.

Ariba Analysis (Spend Visibility), Ariba Buyer, Ariba Contract Management (ACM), and Ariba Sourcing and Ariba Contract Workbench are trademarks or service marks of Ariba, Inc.