[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Ganymede dev2b released
|
> -----Original Message-----
> From: Jonathan Abbey [SMTP:jonabbey@arlut.utexas.edu]
> Sent: Monday, May 04, 1998 5:00 PM
> To: King, Curtis
> Cc: jonabbey@arlut.utexas.edu; mulvaney@arlut.utexas.edu
> Subject: RE: [Ganymede] Ganymede dev2b released
>
> | Thanks for the quick response. Overall I found ganymede an
> excellent
> | tool. I was going to build something similar using Oracle and Java,
> but
> | this has saved me a huge amount of time. I like the customization
> | ganymede affords. I am looking forward to the first production
> release.
>
> Thanks for your bug report! We unfortunately have not been getting a
> great
> deal of feedback on Ganymede so far, outside of the lab here. Every
> little
> bit helps.
>
> I'm very very interested in hearing how the customization goes for
> you..
> I'm trying to make the plug-in classes as powerful and as customizable
> as
> I can, but I am quite sure that there are things that are not well
> documented
> at this point, and extension hooks that need to be added. Have you
> done
> any plug-in coding, or have you just been working with the schema
> editor
> at this point?
>
[ck] I have done about 50-50 of the work with the schema editor and the
plug-in classes. Most of this work is to enforce the use of corporate
ID standard and change the NIS maps created to meet our needs.
> How big are your user_info and hosts_info files in GASH? How have you
> found the performance of the Ganymede server and/or client? What sort
> of machines are you running things on?
>
[ck] We have 26 users and 10 groups in GASH. This will grow to about
800 - 1000 users, 40 groups, and 100 hosts once Ganymede is implemented.
The server is a Sparc 10 with 128 Megs of RAM, the clients are a Sparc
450 back to a number of xterms, also a number of NT desktops running the
client. The performance has been good as the NT tools we use for the
same function as Ganymede. Better than Tivoli which we use to add users
to most Unix servers currently. We have about 80 Unix servers for
database, file/print and each users is added to each server as needed.
Needless to say a huge man power overhead and a big security problem.
So, I used GASH to demonstrate how to better manage users when I started
working here a year ago. But, the security group which manages user ids
wanted a GUI which Ganymede gives us and looks like the NT tools. The
security group has been very impressed with Ganymede. Also, I want to
manage DNS with Ganymede because I am still having problems getting the
guys to remember to add arpa entries. Which shows I am not that good of
a teacher :).
> | I am looking forward to the first production release.
>
> I think we're really getting close to it, at least at the level of
> GASH.
> Here at the lab we're trying to evolve the system/network schema to
> support arbitrary subnetwork allocations, the ability to provide a
> subnet
> of 8 or 16 addresses for our ISDN home users, and the ability to
> support
> an indeterminate amount of DNS domains for use with an ISP-type
> environment.
>
> All of which is to say that what defines a production release varies
> according
> to the needs of the adopter.. the passwd/group support would probably
> be quite
> usable right now, assuming that the adopter wrote up their own
> nisbuilder
> script.
>
[ck] I agree that the NIS support is usable now. Will this part of the
schema change in future releases? I am under pressure to roll out this
part. BTW, I am going to load the database to where I am thinking it
will grow too for performance testing. I will let you know how Ganymede
performs.
> | Thanks again,
>
> And thank you.
>
> | ck
>
> ----------------------------------------------------------------------
> ---------
> Jonathan Abbey
> jonabbey@arlut.utexas.edu
> Applied Research Laboratories The University of Texas
> at Austin
> ----------------------------------------------------------------------
> ---------