I just did this about a month ago... first the "why": My single TSM database had grown to its architectural limit: 530GB. I had no choice other than splitting TSM into multiple instances (well, I could have "frozen" the big TSM and started over... and we actually ended up doing that also... but multiple instances seemed like a good approach for a number of reasons). We didn't have the budget for new servers, and our TSM server is pretty hefty - an HP rp7410 w/ 8 CPUs and 12GB RAM. All DB and log volumes are on an HP XP512 (Hitachi Lightning).
Now the "how": Use the same TSM install directory as per the install guide. Use the same host name/IP address for each instance, but assign a unique set of ports (TCPPort, HTTPPort, SHMPort) for each instance. Then assign the appropriate port in the dsm.sys/dsm.opt in each client. And of course register the clients in the corresponding TSM instance. As far as tape access, using one server means you only need to have the drives physically connected to that server (unless you're doing LAN free backups)... and this may enable sharing your SCSI connected drives, although I haven't tried that... we have two SCSI connected DLT drives, but we almost never use them anymore. One other recommendation, which I did... do not run all of your TSM instances as root (you are on Unix , right?). Doing so, if you do a "ps -ef|grep dsmserv" command, you can't tell which process belongs to which server without doing more research .. looking up the PIDs. Instead, create a unique user for each TSM and run it under that user. Make sure you "chown" all of your config files and DB/log/stgpool volumes to that user. This will also prevent the wrong TSM from accessing another TSM's storage... this can be easy to do if you use raw logical volumes as we do. And now, your ps output will show the user for each dsmserv. It's actually quite easy to set up an additional instance once you get all the steps down... I'll share my approach in more detail if anyone is interested... Regards, Robin Sharpe Berlex Labs "ADSM: Dist Stor Manager" <ADSM-L@VM.MARIST.EDU> wrote on 05/22/2006 05:28:32 PM: > Other than the obvious hardware cost savings, I don't really see the > advantage of multiple instances on the same hardware. (I haven't > decided yet if we would use one beefy server or two medium servers.) If > you load up multiple instances on the same server, do you give them > different IP interfaces to make distinguishing between them in client > configs and administration tools easier? Tape access-wise, is there a > hardware advantage putting multiple instances on the same system? > > Any recommendations on any of this? Your help is appreciated. > > Dave >
