How to fix a broken Data Lineage in TIBCO Data Virtualization?

How to fix a broken Data Lineage in TIBCO Data Virtualization?

book

Article ID: KB0072307

calendar_today

Updated On:

Products Versions
TIBCO Data Virtualization 8.2 and earlier

Description

When working with TIBCO Data Virtualization, you 'may' find the lineage for some database objects to be broken for one TDV instance while other instance in the Cluster display  the complete lineage.

Here is a screenshot for reference:

User-added image
This article explains how to resolve this issue. 

Issue/Introduction

How to fix a broken Data Lineage in TIBCO Data Virtualization?

Environment

All supported Operating Systems

Resolution

This issue 'may' happen due to a recent metadata corruption. Here are few options to resolve this issue. 

Note: If Option I and II do not help, implement the solution from Option III. 

Option I:
(1) In a Cluster scenario, enable the configuration 'Copy Repository Database For Cluster Join' and then remove the nodes from the cluster. This configuration can be found in TDV Studio > Administration > Configuration > Server > Configuration > Cluster.  
(2) Re-import the Full-Server Backup in the Timekeeper node.
If the timekeeper node shows the correct lineage, other nodes can be directly added to the timekeeper node. 


Option II: 
If the issue pertains for few Views :
(1) Open the SQL panel of the View and perform an alteration by simply adding some blank spaces in the View SQL. Save the changes. 
(2) Re-open the View tab and create the lineage. 


Option III:
(1) Take a Full server backup from TDV Studio. (Administration tab -> Full Server Backup). Verify the backup is valid.
Additionally, take a backup of the existing metadata by keeping a copy of TDV Server install directory.  
(2) If environment is part of an Active Cluster, take the nodes out of the cluster (TDV Web Manager -> Configuration -> Cluster) and now disable the configuration 'Copy Repository Database For Cluster Join' for all the nodes.
(3) Stop the TDV Server for one node, but keep the TDV repository running. Close TDV Studio.
(4) Drop the schema using repo_util.bat (Windows) or repo_util.sh (Linux) located in the [TDV_Server_Install_dir]\bin folder.
Example for reference:  
(i) List files under <TDV_INSTALL>\conf\repository\postgresql beginning with 'composite_clean'.
dir D:\TIBCO_DV\conf\repository\postgresql\composite_clean*
(ii) If the list contains files named composite_clean-nnnnnnnnn.sql ('nnnnnnnnn' represents a version number), run repo_util.bat with the arguments -dropSchema -schemaDropScript <highest versioned file>
For instance if the list is:
        -  composite_clean-080200001.sql
        -  composite_clean-080401001.sql
        -  composite_clean.sql     
Then composite_clean-080401001.sql is the highest versioned file, and the command is:
repo_util.bat -dropSchema -schemaDropScript D:\TIBCO_DV\conf\repository\postgresql\composite_clean-080401001.sql
(iii) If the list does not contain any versioned files (i.e. only composite_clean.sql is present), the schemaDropScript argument is not required and the command is:
repo_util.bat -dropSchema 
 Output:

register basic services   Connecting to database...      You are about to permanently delete all metadata in the server's   repository database. All server resources including data sources, views,   scripts, and definitions will be deleted. This action cannot be undone. 
Do you want to proceed?
[y,n]: y   Dropping schema...   The Composite server repository schema has been dropped.

(5) Start TDV Server and verify below entry in cs_server.log.
'Server started with an empty database.  Creating the initial system database entries.'
Now, the TDV instance will look like a fresh installation.
(6) Import the Full Server backup from TDV Studio.
Verify the complete lineage is displayed for all database objects.
(7) Create a cluster on this node (this node will be treated as the TimeKeeper) and add remaining nodes to the cluster.