Something wrong with the anoncvs servers?

[Mark Hamstra <mhamstra sullivan bentley com>]

Something seems to be badly out of sync with anoncvs.  As an example, I
can generate the following results from identical cvs commands issued
about 30 seconds apart:

*** First, what looks to be a normal log...
***

  cvs log libgnome/gnome-magic.c

  RCS file: /cvs/gnome/gnome-libs/libgnome/gnome-magic.c,v
  Working file: libgnome/gnome-magic.c
  head: 1.5
  branch:
  locks: strict
  access list:
  symbolic names:
  keyword substitution: kv
  total revisions: 5;     selected revisions: 5
  description:
  ----------------------------
  revision 1.5
  date: 1998/12/09 01:45:53;  author: hvr;  state: Exp;  lines: +5 -5
  a few warning/code-cosmetic fixes
  ----------------------------
  revision 1.4
  date: 1998/12/09 00:40:59;  author: unammx;  state: Exp;  lines: +9 -0
  Add James Youngman
  ----------------------------
  revision 1.3
  date: 1998/12/08 23:47:25;  author: drmike;  state: Exp;  lines: +1 -1
  Changed temporary variable to an int from a short, otherwise we
  seem to overwrite the stack if the fmtstr is a "%o" (octal) on my
  box. Thanks to Owen for seeing this bug.

  Dr Mike <drmike@redhat.com>
  ----------------------------
  revision 1.2
  date: 1998/12/08 20:22:56;  author: sopwith;  state: Exp;  lines: +64
-11

  gnome-magic now precompiles mime-magic.
  Start work on mimifying mime-magic.
  ----------------------------
  revision 1.1
  date: 1998/12/04 00:13:41;  author: sopwith;  state: Exp;

  Little bit of work on detecting mime types by magic. Don't use (yet).

=============================================================================

*** Followed by a truncated log...
***

  cvs log libgnome/gnome-magic.c

  RCS file: /cvs/gnome/gnome-libs/libgnome/gnome-magic.c,v
  Working file: libgnome/gnome-magic.c
  head: 1.1
  branch:
  locks: strict
  access list:
  symbolic names:
  keyword substitution: kv
  total revisions: 1;     selected revisions: 1
  description:
  ----------------------------
  revision 1.1
  date: 1998/12/04 00:13:41;  author: sopwith;  state: Exp;

  Little bit of work on detecting mime types by magic. Don't use (yet).

=============================================================================

I see similar effect when trying to update source; i.e., constant fights
as to what the current state of the archive is, resulting in large
numbers of patches/updates between at least two different states.  It
looks like the different DNS balanced cvs servers (they are being round
robin DNS'd, yes?) are pointing at different repositories or something.
Needless to say, this makes source maintenance on my end something of a
probabilistic and frustrating affair.
 
Does anyone else see similar effects, know what is going on. or --more
importantly-- know how to fix/avoid the problem?
 
--
Mark Hamstra
Bentley Systems, Inc.