[OpenAFS] Rehashing the prospective death of uss

Russ Allbery rra@stanford.edu
Tue, 31 Aug 2010 17:49:27 -0700

Jason Edgecombe <jason@rampaginggeek.com> writes:

> Was a consensus reached on whether or not uss should be kept around? We
> still use it at our site with the -skipauth option to avoid any kaserver
> stuff. We create the Krb5 principals with a perl script run via remctl.

We decided not to kill it, but as yet no one has had the time to work on
improving it.  My personal feeling is that the best way forward would be
to surround all the kaserver bits with some sort of #ifdef so that as soon
as we pull kaserver from the tree, we can easily remove all of that code,
and then update the documentation to stress -skipauth.  For extra bonus
points, one could add kadmin support to uss, which would enable it to work
as it did before.

My personal feeling from looking at uss is that it's something that we've
reinvented at Stanford with a lot of additional capabilities, and if we
were designing it again, I think it could be made much better.  But until
someone has time to do that work, if it's useful, I'd rather not remove

Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>