Is this Bladelogic installed on Physical Box or VM? If it is VM I would recomend you to take snapshot.
If it is installed on Physical Box, then take Back-up of File Server, script folder located under NSH DIR and Database. This is Sufficient to restore.
Is your question answered?
What I do: directly before upgrading, when all server processes are stopped, I make a complete copy of exactly the directories you mentioned:
e.g.: cp -rp /opt/bmc/bladelogic /opt/bmc/bladelogic.8.2.01 (for your version)
And the same way for /usr/lib/rsc -> /usr/lib/rsc.8.2.01
This way you could - in case of some severe failure - just rename the directories and restart your server with no addtionl time lost. In the original directories I delete the logfiles (appserver-log, spawner-log and so on) to save space - they are in the copy!
One thing you didn't mention are the startup scripts /etc/init.d: i.e. blappserv, blprocserv, blpxe, bltftp (both if you do provisioning only) and rscd.
There are also some symbolic links in /lib/ or /usr/lib (I just don't remember and no environment at hand), maybe /something like /lib64.
if you do a "locate bl | grep lib" you should get them (unfortunately not only them...).
don't forget to backup the database.
But here is an order : backup database must be newer. The depot folders must be older.
Good point! I think if first stopping BL - including the fileserver agent - and doing the filesystem-backup will leave the database in a suitable state.