>>>>> "Greg" == Greg Troxel <gdt_at_ir.bbn.com> writes:

  Not that I think it makes much difference but I use tunnel mode for my
  ipsec not transport mode.

    Greg> I suspect the IPsec part is not relevant,

Isn't there a problem where IPsec and fragmented packets mix very
poorly because the fragment information is encrypted?

Of course, I first heard about this from traffic on this list, in
particular some packet size was decreased to alleviate such a problem
(IIRC the expert who identified this was you, Greg).  But I've since
done some studying and it seems to me that the Coda protocol is
complex enough that another such problem might be lurking.

Re the conflicts themselves:

I've also seen wedgedness in the conflict resolver, more or less like
what you say.  I think I've mentioned it before.  But my normal
response to conflicts anymore is venus -init because the content
that's getting hosed is a fairly pristine CVS tree, so it's cheap and
fast (and usually I'm in a hurry to get to the data ;).

Another thing that I've seen a lot is a situation where the directory
claims to be in conflict, but the repair tool can't find any
conflicts.  These always involve a typical CVS rename - create -
rename sequence for replacing a file.

I'm not sure I'm seeing disconnection, when I get back I'm connected
(these heavy updates take place at 4am).  However, IIRC there are
several different trees being cvs updated, and the early one warns
about a time out, the later ones say "can't create" all the CVS
Entries.backup files -- could this be a disconnect / lose token /
reconnect kind of thing?  What do I look for in the Coda logs?

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
 My nostalgia for Icon makes me forget about any of the bad things.  I don't
have much nostalgia for Perl, so its faults I remember.  Scott Gilbert c.l.py