I have been working on some refreshed versions of Coda Documentation.  You
can find a build process for a Coda server and client from scratch at:

http://info.pittsburgh.intel-research.net/project/isr/coda/chapter06a.html

and

http://info.pittsburgh.intel-research.net/project/isr/coda/chapter06b.html

I don't claim that they are complete, as they are (always) a work in
progress, but they are much more accurate than what is generally
available.  It goes into detail for setting up a Coda installation on
RedHat systems (2.4.xx kenrels), including the little "gotchas" that are
normally not mentioned.  Though, I imagine that my documents are
applicable to any unix/linux style system/distro.

Hope this helps some folks.

Casey Helfrich
Intel Research Pittsburgh

*********************************************************************
*                          Casey Helfrich                           *
*                        cjh2_at_andrew.cmu.edu                        *
*********************************************************************

On Tue, 20 Apr 2004, Ivan Popov wrote:

> Hello Burt,
>
> > A problem is that the documents I am using are the Administrator's Manual
> > from 2000 and the HowTo from 1999.
>
> unfortunately, I think there is no better documentation.
>
> The most reliable and up-to-date pieces of information are present in the FAQ
> section and in the list archive (but you have to know what to look for :)
>
> > >given a lot of care and education,
> > >and certainly a lot of your time helping them when something goes wrong.
> >
> > After years of using AFS, I never managed to solve the ".: not found"
> > (something like that) problem, where parts of the filesystem disappear.  I
>
> Hmm, I'm rather used to ".: permission denied"
> as I do not use a proper pam module for my screensaver.
> But it is just a minor annoyance forcing you to clog explicitely.
> I suppose you mean real problems like when the cache manager dies.
> Then of course it is a bigger trouble.
>
> By the way, talking about friendliness - Jan, I want to bother you again
> with that idea - let Venus use cached access rights on cached objects
> until a valid token is present...
> It would dramatically improve "friendliness" of the file system, without
> sacrificing security. Coda does have support for the disconnected mode -
> that's why we can both support network credentials expiration
> and behave user friendly.
>
> > ran into that same thing with coda when I made that mistake of accessing
> > the same thing through 2 different paths. Usually it required restarting
>
> A bit strange, it should not cause something like that - but now it is
> hard to know what happened.
>
> Regards,
> --
> Ivan
>
>
>
>