[1899] in Release_Engineering
Re: Staff Disk Space document
daemon@ATHENA.MIT.EDU (Ken Raeburn)
Mon Nov 6 11:41:18 1989
Date: Mon, 6 Nov 89 11:40:14 -0500
From: Ken Raeburn <Raeburn@MIT.EDU>
To: roden@ATHENA.MIT.EDU, vice-squad@ATHENA.MIT.EDU, rel-eng@ATHENA.MIT.EDU,
In-Reply-To: Jon A. Rochlis's message of Sat, 4 Nov 89 17:13:53 -0500,
In addition to those mentioned in Jon's comments, I think there are
some other issues that haven't been addressed.
There are services other than file service provided by some of the
machines in question; none of these have been addressed. If the last
paragraph in the "current method" section is modified to read "...no
longer be practical to have each group allocated a dedicated file
server", thereby keeping other server issues open, this can be
sufficiently sidestepped. (The alternative is to address other
services in the same proposal.)
The next paragraph says:
.... This move would not be without inherent benefits *and*
risks. I will attempt to outline these in the following
paragraphs.
Maybe I missed it, but I didn't see much discussion of the risks of
this proposal. I would count among them the apparently increased
sensitivity to network failures, and the problems recovering from such
failures. (Having to reboot is not the ideal solution.) AFS is also
rather flaky on the PMAX at times, I might mention. Then there are
whatever differences are inherent in AFS when it is working
correctly....
With regard to administrative access to servers: The issue of "status"
is mentioned, but it isn't clear what (if any) implication this has.
-- Ken
