I want to thank everyone for their responsses so far - keep 'em coming though.
First of all let me respond to Jeff's inquiry, "Is there something in
particular that's got you irritated enough to migrate an entire enterprise to
a different software?"
A lot of things. Here's a list of just a few (from the Sun side of things):
1. With over 200 clients, updating them from one release to the next requires
going to each client (via login of some kind) and doing an interactive
pkgadd. I'm told that NBU allows client updates to be "pushed out." This
would save a lot of time, and be version reliable.
2. Version control is an issue with LGTO - if you get a patch for something,
there is not centrally located way to track which box has what patch, and
it is usually not in pkg format. LGTO just doesn't track version changes,
such as date and time and version.
3. Scripting *can* be a nightmare - while some folks love mminfo, it takes a
long time to get the incantations correct, and nsradmin -i is a joke, and
not easily scripted. There may be a perl module out there with bells and
whistles, but I've yet to find it.
4. Size limitations on directives.
5. Installing devices and jukeboxes are a nightmare - if you mistype something
in during jbconfig, you must start all over again from the beginning.
LGTO doesn't support drive serialization.
6. LGTO doesn't support MacOS clients, NBU has support.
7. LGTO cannot stage to disk w/o purchasing an option, NBU can stage, need
other products to to automatic staging, however.
8. LGTO writes in proprietary format to tapes, NBU is modified gnu-tar.
9. Takes forever to pre-label legato tapes PRIOR to backup. NBU does it at
backup time (less SA intervention).
As for NDMP - that is something I am interested in - however, if NBU has lousy
support for that, we'd most likely use LGTO.
I'd be interested in what the NBU issues are with NDMP. I didn't know that
there ARE issues in this area!!
Comments, Suggestion, Jokes ?? :-)
Thanks, guys,
deb
|
