Yep... it's possible to arrange a single-drive environment; it's highly
dependent on the onsite person doing tape mounts on request, and organizing
the tape pool to satisfy your retention requirements -- if you're lucky, get
sufficient disk pool storage to allow either single-drive reclamation or
copy-pool support.
I've had two different situations, we resolved them as follows:
1. Manual single-drive -- a file server with Exchange server; we
configured it for Incremental forever (the file server node), daily
incremental with weekly full for the Exchange; we installed admin client on
the desktops of 2 or 3 onsite supervisors, so they would watch for tape
mounts, and react accordingly. We used server-to-server across a WAN for
the database backups; again, daily incremental, weekly full (on the
weekend, when traffic was low). This is probably the worst scenario for
trying to use TSM -- but, that's what we did; it was running for about a
year when I left the account -- turned it over to centralized backup
personnel, who continue the operation. This was a "test" to see how good or
bad it would be to manage remotely; it was "okay", but one must always take
initiative to monitor the onsite folks performing tape mounts. Rather than
try to do reclamation, we did once-a-month "Full" backups (by changing the
policyset to absolute for one cycle), which would cause older tapes to
(ultimately) go to "Empty", since retention of Exchange was one-month, files
was 30-45 days... poor man's "trick" reclamation.
Since this was a DLT, we minimized tape thrashing by running both daily
backups in close succession, using same media for all the data... used one
tape per week, accumulating approx. 15 tapes total, as they recycled.
Refreshed all new tapes once per year. Another reason this is not quite as
cool as we'd like is it lacks copy-pool protection against media failure --
for case of a large restore request (as in a disk drive failure). I think,
more likely, a simpler approach, would be better -- weekly-fulls might be
better than the monthly-absolute -- in case of media failure, you'd still
have another copy of all but the most recent week's data).
2. Single-drive library with 6 or 7 slots -- installed sufficient disk pool
to handle single-drive reclamation, as per Admin. Guide; works fine,
allowed 14-day PIT supporting one tape drive's worth of file-served data
(approx. 65 GB in this case); that, plus diskpool used for storing primary
copy (never migrated), tape slots used for two copy pools (one kept onsite,
the other sent offsite) plus db backups. 14-days of PIT translates to
approx. 1.7 x file-server-capacity (using 5% per day data change/add/delete,
rule-of-thumb). This is the minimal, workable configuration for TSM (imho);
minimum recommended configuration would be TWO tape drive library with
sufficient slots to hold all backup data for desired number of versions,
plus some slots for db backups & daily copy pool tapes... can I do daily
incremental db-backups to a single output tape (appending each day's
data)?!? Never did reclamation -- all unexpired backup data kept in disk
pool, never migrated... eventually, MOVE DATA could be used to vacate disk
pool to allow for single drive reclamation of the onsite copy pool.
Hope all this helps!
Regards,
Don
|
