I recently did a network based upgrade of a 270GB DB. Previously TSM could take days to due the upgrade of a DB so large, but my experience with performing upgrades for other companies and some suggestions from my Tivoli Consultant had me convinced it could be done in under 24 hrs. The network method is kind of a misnomer here since we used the loop-back address, so the upgrade was done in place on the server. After allocating additional disk space (a lot of disk space) and defining the user ID TSM would run under, I started the upgrade process by running the dsmupgrd preparedb process. The upgrade took less than an hour and completed successfully. I copied our devconfig and volhist files to an alternate location and then started the insert under the id of the new TSM 6.2.4 instance. I then switched back to root, cd'd to the upgrade directory, set my environmental variables accordingly, and then started the extract.
The extract took 5 hours and ran without issue. The insert ran for 11 hours and completed without errors. The overall time from start of the upgrade to end was 13 hours. My reason for testing was to verify the time frame needed so the applications that rely on TSM for log offloading could add to their additional log space ahead of time. If anyone has suggestions on how I can make the extract/insert performance even better feel free to post a comment.