Recommended method of moving the existing CLM environment on Windows to Linux

If you have a common public URI root (e.g. https://myserver/) than this can be used by the proxy. You don't have to do a server rename, in fact it would be the wrong thing to do.

The applications keep the public URI, they just run on a different machine. Since the proxy forwards all requests they simply are not aware. See the pictures in https://jazz.net/wiki/bin/view/Deployment/RecommendedCLMDeploymentTopologies5 . It is important that the applications think they have the same public URI that is used by the proxy.

If the applications don't have a common public URI root (the same hostname/port prefix) you would have to do a server rename to get a common one.

In general the approach seems to be OK.

There are other files you might want to consider too. See https://jazz.net/wiki/bin/view/Deployment/BackupCLM

Hi Ralph

 

Thanks. I have created a staging environment for CLM 4.0.3 on Linux just to replicate their current production environment on Windows ( CLM 4.0.3, RRDI 2.0.x, DB2 9.7 and WAS 7)

 

Here is what I have tried:

1. Export the CLM DB2 9.7 database on Windows using Db2move “database name” export –l “<path to the backup>\CCM\lob”, for database JTS, CCM, QM and DW
2. Created SQL script for DDLs and tables, “db2look -d mydb -e –l –x -o mydb_create.sql”
3. On the Linux machine installed DB2 10.5 and created databases for JTS, QM, CCM and DW
4. created tables and schemas using “db2 –tvf  “mydb_create.sql”
5. Imported the data using “db2move db import”. Had issues with DW import as there were lot of dependent tables. So ran Load instead of import and set integrity for all the tables using a .sql script
6. Installed WAS 8.5.5 on the proposed CLM 4.0.3 server on Linux and created a profile.
7. configured LDAP
8. Installed CLM 4.0.3
9. Deleted the “conf” folder of staging environment and replaced it with the “conf” from the Windows CLM 4.0.3 production environment
10. Modified the teamserver.properties files with new entries matching the locations of Databases, index location, base url and data base folder location of the Linux staging environment
11. Ran the clm_was_config.py Jython script to configure WAS profile
12. Manually checked if all the configuration are correct: In JVM custom properties
13. Restarted the WAS profile
14. Deployed the CLM WAR files using clm_deploy.py
15. Configured security roles to user mapping for ccm, jts and QM
16. added and entry in /etc/hosts file with the host name of the production Windows machine and ip address of the current Linux machine
17. Accessed the original CLM production URL from the client machine and could login to the new CLM 4.0.3 Linux environment. Got an error regarding RDF index.
18. Stopped ServerRebuild the databases indices as well as the Text indices for JTS, CCM and QM with repootools.
19. Restarted the server and could log back in and see all the data.

 

I am now in the processes of upgrading the CLM 4.0.3 on Linux to CLM 5.0.1 (using the CLM interactive upgrade http://www-01.ibm.com/support/knowledgecenter/SSYMRC_5.0.1/com.ibm.jazz.install.doc/topics/roadmap_clm_upgrade.html?lang=en ) and then make it to a full distributed environment with reverse proxy.

 

I presume I need to do a server rename after upgrading to CLM 5.0.1 to make it to a fully distributed environment. Please correct me if I am wrong or anything I need to consider before this move. Thanks

 

Best Regards

Abish M Zachariah

Hi Ralph Thanks. I have created a staging environment for CLM 4.0.3 on Linux just to replicate their current production environment on Windows ( CLM 4.0.3, RRDI 2.0.x, DB2 9.7 and WAS 7) Here is what I have tried: 1.Export the CLM DB2 9.7 database on Windows using Db2move “database name” export –l “\CCM\lob”, for database JTS, CCM, QM and DW 2.Created SQL script for DDLs and tables, “db2look -d mydb -e –l –x -o mydb_create.sql” 3.On the Linux machine installed DB2 10.5 and created databases for JTS, QM, CCM and DW 4.created tables and schemas using “db2 –tvf “mydb_create.sql” 5.Imported the data using “db2move db import”. Had issues with DW import as there were lot of dependent tables. So ran Load instead of import and set integrity for all the tables using a .sql script 6.Installed WAS 8.5.5 on the proposed CLM 4.0.3 server on Linux and created a profile. 7.configured LDAP 8.Installed CLM 4.0.3 9.Deleted the “conf” folder of staging environment and replaced it with the “conf” from the Windows CLM 4.0.3 production environment 10.Modified the teamserver.properties files with new entries matching the locations of Databases, index location, base url and data base folder location of the Linux staging environment 11.Ran the clm_was_config.py Jython script to configure WAS profile 12.Manually checked if all the configuration are correct: In JVM custom properties 13.Restarted the WAS profile 14.Deployed the CLM WAR files using clm_deploy.py 15.Configured security roles to user mapping for ccm, jts and QM 16.added and entry in /etc/hosts file with the host name of the production Windows machine and ip address of the current Linux machine 17.Accessed the original CLM production URL from the client machine and could login to the new CLM 4.0.3 Linux environment. Got an error regarding RDF index. 18.Stopped ServerRebuild the databases indices as well as the Text indices for JTS, CCM and QM with repootools. 19.Restarted the server and could log back in and see all the data. I am now in the processes of upgrading the CLM 4.0.3 on Linux to CLM 5.0.1 (using the CLM interactive upgrade http://www-01.ibm.com/support/knowledgecenter/SSYMRC_5.0.1/com.ibm.jazz.install.doc/topics/roadmap_clm_upgrade.html?lang=en ) and then make it to a full distributed environment with reverse proxy. I presume I need to do a server rename after upgrading to CLM 5.0.1 to make it to a fully distributed environment. Please correct me if I am wrong or anything I need to consider before this move. Thanks Best Regards Abish M Zachariah