How is the Jazz data warehouse migrated from vendor DB to another?
We are running CLM 5.0.2 iFix15 with RTC and DNG on linux under WAS 8.5 with a DB2 DB. We are migrating our DBs from DB2 to Oracle.
I have found the documentation on how migrate DB vendors usinge the repotools-<APP> -export( and -import) commands to generate the ".tar" files for each of the application databases, but I don't see or understand how the data warehouse data is handled in this process. Each of my application "teamserver.properties" files has a jdbc connection to the specific application database, but each also contains a jdbc connection to the common data warehouse database.
So following this overly simplistic example, would expect:
repotools-jts -export to export just the JTSs application data from the database defined at:com.ibm.team.repository.db.jdbc.location=
repotools-ccm -export to export just the RTC application data from the database defined at:com.ibm.team.repository.db.jdbc.location=
repotools-rmts -export to export just the RRC/DNG application data from the database defined at:com.ibm.team.repository.db.jdbc.location=
Can someone tell me how the data warehouse is migrated or what command(s) are used for export/import for the data warehouse (defined at: com.ibm.team.datawarehouse.db.jdbc.location=)?
Thanks, Steve
4 answers
As far as I can tell, there is no repotools export for the data warehouse. I think you will have to create with a new database and the data warehouse will have to be recreated from the application data. There might be some data e.g. for build results that are already deleted, that might not be possible to recreate.
Hello Steve,
Ralph is correct. You will need to recreate the data warehouse from scratch on the new DBMS.
On the topic of what you will lose, all the historical trends and metrics cannot be recreated.
So you will have the reports based on metrics only showing data from the day the first ETL on the new data warehouse is run onward.
For example this means the Burnup/Burndown reports will be broken for all the sprints in progress.
They will start again to work as intended for the sprints created after the data warehouse change.
Best Regards,
Francesco Chiossi
Hello Kevin,
the data warehouse structure between Oracle and DB2 is very different, so I don't expect a generic migration tool to work properly.
You can see more details on how the 2 data warehouse are created in these wiki pages:
Oracle: https://jazz.net/wiki/bin/view/Main/MoreControlOverTheOracleDataWarehouseSetup
DB2: https://jazz.net/wiki/bin/view/Main/MoreControlOverTheDB2DataWarehouseSetup
In the Oracle data warehouse the DB2 schemas are implemented as inactive DB users owning the objects.
Best Regards
Francesco Chiossi
See: http://www.oracle.com/technetwork/database/migration/db2-084087.html
I'd be reluctant to toss aside historical warehouse data if it is even slightly possible to migrate.
Comments
Kevin Ramer
May 22 '17, 4:45 p.m.You can do db2move export to get the tables into a collection of ixf files. I don't know if Oracle can read'm though.