I'm in the process of moving from a server that has just developed on a
Risc J40 to a well-planned (I hope) server on a SP2 node, starting with
those machines our department is responsible for.  I'm using export/import
and have some questions about results.

1.      As I posted a couple of weeks ago when I started, a syntactically
correct import kept failing.  It took support two weeks to tell me I
couldn't import a client into a domain different on the new server from the
old (I resolved it by creating a dummy domain on the new machine, doing the
import, and updating the client to the desired domain).  Is this a feature
or a bug?

2.      Some days I don't get the import done before the baseline backup is done
on the new server.  When TSM finds a duplicate name (file?, directory?
filespace?), it creates a new filespace to which it gives a similar
name.  Will these new filespaces be treated like the real ones on the
client when it comes to versioning, inactive retention, etc?  If an db
entry for client\c: expires, will the corresponding db entry fro client\c1:
expire with it?

3.      Some of these have large archives attached.  Since I only have two
drives available at the present, TSM is happier if I import twice, first
with filedata=archive and then with filedata=backup, so that migration can
run while the import is in process.  The import archive process creates a
filespace name for every filespace on the client, whether it has archived
files or not.  If I'm not too busy to remember, I delete the unused
filespaces before starting the import backup process and everything looks
normal to the user from the dsmc command line or the gui or the web
client.  If I do forget, I have a client with duplicate filespace names for
every filespace on the machine, each with the identical characteristics on
the q fi display.  In this case, the import seems to have landed in the
generated filespace.  Can I safely delete the original filespace, e.g.,
/var, and rename /va1 as /var, etc?

4.      Combining the previous two, if I find myself with duplicate filespaces,
but an incremental has been done in the meantime, where do those backups
go, /var or /va1?  If I wind up with two filespaces being used, does my
previous question about expiration apply? or do the incrementals only go
into one filespace?  What effect does this have on PIT restores?  Do I have
to teach admins to look in multiple places for files for an unknown length
of time?

5.      Depending on answers to the preceding, should I start over?

Thanks.


Fred Johanson
System Administrator, ADSM
S.E.A.
University of Chicago
773-702-8464

Reply via email to