On Thu, 19 Oct 1995, Mark M. Klivansky wrote:
> I would like to ask ADSM users if many are happy with a product or support.
This question has got mainly positive answers on this list which squares
well with the observation that most list participants have a mainframe
based ADSM installation for backing up real small systems. We, however,
use ADSM for a couple of larger workstation clusters and are less than
happy with it:
There is no usable tool for salvaging the ADSM database. AUDIT DB,
tried once, had to be aborted after five weeks without result. We know
of other sites with large numbers of files who, too, are very alarmed by
the lack of this feature. Running ADSM for more than a million files or so
seems to be extremely risky.
There is no support of clusters. The concept of "node" is inadequate to
describe the common data of a workstation cluster. In our case, we have
selected to have each client work for two nodes: one is its private node
and the other is a node name common to the entire cluster. Otherwise it
would not be possible to reload data from another workstation than the
one from which it was stored.
ADSM does not support AFS or DFS; in the contrary, ADSM's security
mechanisms undermine the security provided by Kerberos.
The GUI is inadequate for workstations with large directory trees.
Menus take a quarter of an hour to come up.
The whole thing seems not to be designed for client/server environments,
but for individual PCs and workstations. I have written a lengthy paper
on this subject (already distributed to this list).
As to support, we have as a major problem that errors are discussed
instead of fixed. My favorite example is this one:
The ADSM client on AIX, even that one that allegedly supports AFS, is
unable to detect mount points. This may lead to a situation that a backup
includes worldwide data if the user chose to mount a volume from a distant
computer (which he, other than under NFS, is free to do). This error has
already prevented full usage of ADSM at at least two other sites before we
purchased ADSM. At that time, we expected the error to be fixed,
especially since it is probably quite easy. We reported the error in
February as a PMR which was closed shortly after without a fix. We
discussed the problem in May or June with the sales people who wanted to
prevent it from appearing in a list of errors to be fixed in the warranty
time of nine months ending about March 1996. We discussed it with
Lab-Link, Geneva, who sided us. During all that time I have written
numerous letters explaining why I feel that an error should be fixed
instead of discussed.
The GUI problem (see above), probably even easier to solve, has the same
history including the fact that the PMR has disappeared without being solved.
The more we get experience with the "support" process, the more do we
understand those from other sites that have resigned and now save the
effort of writing PMRs.
> I have eggs on my face for recommending implementing ADSM.
Fortunately, ADSM took six months to become operational. So, when the
whole thing crashed and had to be rescued by deleting most of the data,
there were not too many users involved. The number of eggs in my face is
therefore limited.
Since the installation of ADSM in January, I have had one person work
exclusively on inventing work-arounds for ADSM's bugs and shortcomings.
Thus, despite its moderate price, ADSM has turned out to be one of our
most expensive products.
===========================================================================
Dr. Helmut Richter
Leibniz-Rechenzentrum X.400: S=Richter;OU=lrz;P=lrz-muenchen;A=d400;C=de
Barer Str. 21 RFC822: Helmut.Richter AT lrz-muenchen DOT de
D-80333 Muenchen Tel.: ++49-89-2105-8785
Germany Fax: ++49-89-2809460
============================================================================
|
