On Mon, Mar 7, 2011 at 12:36 PM, Richard Pieri <richard.pi...@gmail.com> wrote: > On Mar 7, 2011, at 1:25 PM, Jack Coats wrote: >> Even a big data center then was about 4T of disks, and a few thousand >> tapes. Today it doesn't take much >> to get data centers that big in a back room of a mom&pop shop. > > Tangentially, late last week I had to mkfs.ext3 an 8TB file system. Took > about a half hour. That was fun. And scary. At the same time. The > network-based restore from TSM is just slow. > > --Rich P.
Personally I really like TSM, but restores do suck if the files are scattered over all the media in the data center. And due to the way TSM works, it normally is. My favorite way to use TSM was to have enough disk space for a pool or two on disk to keep a copy of everything from the last few days, and then a local tape (or sequential data sets on disk or whatever) for the regular long term backups. And of course have archive copies on tapes (or removable media of whatever kind). These days with multiple TB sata drives being fairly cheap, it would be tempting to use those as the removable media to send off site. When I was working daily with it in Houston, I begged for our DR center to have a full disk copy synced daily from our production TSM server, but it didn't happen while I was there. ... Don't know why I still think and sometimes worry about such things. It is possible to speed up TSM restores, but it takes pre-planning, and probably knowledge of the current versions (the last time I used it was 3 years ago). ... Best of luck Rich ... Jack _______________________________________________ Discuss mailing list Discuss@blu.org http://lists.blu.org/mailman/listinfo/discuss