There is information in the TDP manual. It's not that complicated, except care needs to be taken in configuring the services.
At a high level:
- install the TDP client on local storage of each cluster node.
- place your option file and backup script on a shared resource that fails-over with the SQL instance. This will assure each node has access to it when it owns the SQL instance.
- install a DP scheduler service for each instance of SQL on every cluster node that can own the instance but DO NOT START THE SERVICE (note: you will need to fail the SQL instance over to each node as you are working on it)
- add the scheduler service as a "generic service" cluster resource (if unsure how to do this refer to the manual). I prefer to disable the option on the advanced tab "affect the group". This way if the TSM scheduler service fails it will not cause the cluster to fail-over.
- configure a node on the TSM server. I suggest you use either the virtual cluster server name, or SQL instance name, as to not confuse it with the normal file level backups if performed.
- on each cluster node from a command prompt, change to the TDP install directory and execute "tdpsqlc query tsm" to sync the password in the registry with the TSM server
- from the cluster adminstator mmc fail the cluster over between possible owners to confirm everything is setup correctly, if you experience a failure it is most likely related to the password sync step above.
Here is a link to the manual:
http://www.redbooks.ibm.com/redbooks/pdfs/sg246148.pdf
Hope that helps you out....
~Rick
