Hot Diggety! Prather, Wanda was rumored to have written: > 1) Is there any particular reason to set a max file size for a disk stgpool? > (Assuming a setup where disk stgpool will migrate to tape stgpool) > > >>>The rationale for setting a max file size goes something like this: > > Suppose you have mostly small files backing up, and just a few humongous > ones: say you have a 50 GB disk pool, and one data base that backs up 30 GB > per night.
[...snip great explanation...] Prevent tape thrashing...makes sense. Same reason why I'm not going to enable caching... would speed up restores if I used HSM for certain frequently accessed data or for a big (and recent) restore, but 99.999% of the time, the tape drives are going to be busy writing out tapes. Restores are very rare, so no sense in causing unnecessary head movement/slow media access; I'll just take that extra delay hit at restore time. I know my way around news servers (which is a pretty large 2 TB+ database) and optimize it to prevent disk thrashing, so the idea of preventing tape thrashing (cpu/mem always faster than slow tape access = big hit in perf) makes sense. > 2) Should the TSM server have its own stgpool for backing up itself? > > >>>No need. Ok. > >>> Yes. If you are using the GUI, look at the box at the top of the > window. INCREMENTAL (in most cases use "incremental complete") just backs > up changed files. Pull down ALWAYS BACKUP instead to force a backup > whether changed or not. If you are using the command line, it's "dsmc > incremental" vs. "dsmc selective". Ahh-ha! > Remember though, if your management class/backup copy group sets a limit on > the number of versions you retain for files, that doing forced backups Makes sense. Copy pools used to make copies of whatever primary pool... so if I attach one to all tape stgpools, it will get a copy of data on all the tapes. So in essence, I need 2x number of tapes to cover any given quantity of data; one pile for on-site, and one pile for off-site. That makes more sense. Don't know why that didn't come across so clearly in the documentation. ;) I guess the docs talked more about how to do things rather than what/why, or glossed over the latter. Not uncommon for authors that knows the stuff so well they've forgotten what it looks like to a newbie ;) -Dan Foster IP Systems Engineering (IPSE) Global Crossing Telecommunications
