[151] in SIPB-AFS-requests

home help back first fref pref prev next nref lref last post

Re: aklog in /afs/sipb

daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Sun Jan 14 06:12:59 1990

From: probe@ATHENA.MIT.EDU
Date: Sun, 14 Jan 90 06:13:18 -0500
To: dkk@ATHENA.MIT.EDU
Cc: qjb@ATHENA.MIT.EDU, sipb-afsreq@ATHENA.MIT.EDU, probe@ATHENA.MIT.EDU
In-Reply-To: David K Krikorian's message of Sat, 13 Jan 90 22:36:45 -0500,
Reply-To: Richard Basch   <probe@ATHENA.MIT.EDU>

This discussion has been moved off of vice-squad... it does not concern
most of that list...

  From: dkk@ATHENA.MIT.EDU
  Date: Sat, 13 Jan 90 22:36:45 -0500


  > Right now, there are copies of aklog in serveral places:
  >
  > /afs/athena.mit.edu/service/@sys/aklog
  > /afs/testers.athena.mit.edu/service/@sys/aklog
  > /afs/sipb.mit.edu/service/@sys/aklog
  > /afs/athena.mit.edu/system/afsuser/@sys/aklog

  Kerberos is a Network Services service, and all three of the above
  cells (athena, sipb and testers) use it.  Therefore, it seems to me
  that each cell should have its own copy of aklog.

I, for one, do not think I can keep the testers cell in sync with the
Athena cell, so I have DELETED the copy of "aklog" in the testers cell.

The only reason that copy persisted so long was that the original
announcement of where to find it was in /afs/testers/maint.  If there is
still any dependence on it, I want that dependence broken.  I believe an
early version of "attach" may have also looked in /afs/testers/service
shortly after its renaming.

The only software that I wish to trust is that in
/afs/athena/system/afsuser/${hosttype}bin/ and the copy of "aklog" in
/afs/athena/service/.

I do not care if "aklog" is in each cell personally, but I would
certainly object if people were being pointed to out-of-date software in
other cells.  There have been several changes to various pieces of
software already, not all of which have been publicly announced.

On the flip side, the cell is autonomous, and thus it can be argued that
you should have local binaries.  However, since most clients are in the
"athena" cell, if there is a problem with the primary servers, they are
certainly not going to get to the sipb cell...

On another note, the primary servers for the Athena cell WILL CHANGE.
Its reliance on building 11 will shortly be corrected; there will be a
primary in each of E40, 11, and W20.

Good luck on whatever you decide.

-Richard

home help back first fref pref prev next nref lref last post