[8217] in Athena Bugs
More on "loggin in without a fileserver"
daemon@ATHENA.MIT.EDU (kevles@ATHENA.MIT.EDU)
Mon Sep 16 16:22:47 1991
From: kevles@ATHENA.MIT.EDU
Date: Mon, 16 Sep 91 16:23:07 -0400
To: bugs@ATHENA.MIT.EDU
Cc: kevles@ATHENA.MIT.EDU, timo@ATHENA.MIT.EDU
Hi! Here's some answers from one of the consultants who observed the bug
(that's "timo@athena", aka Tim O'Malley).
1. I am assuming that something is causing the machine to panic. Few
programs call reboot(). There are usually messages which appear when a
machine crashes, and knowing what messages were displayed would help.
He didn't see any messages. He could have missed them, though,
since he was surprized by this behavior, hence not looking in
the right part of the screen.
2. What types of users are reporting this? Is this a problem which
happens to all users on a given fileserver when the fileserver is down,
or is it something which some users see, but not others?
Everyone who reported this had files on Atlas. He thinks Atlas
was the only filesrver down at the time.
3. What exactly is the status of the fileserver when this happens? When
the fileserver is truly *down*, meaning that it will not ping or respond
to any network connections whatsoever, then I doubt that this problem
will occur with high repeatability.
At that point he had not been told that the fileserver was
rebooting (it *had* been all the way down) but about 20 minutes
later he was told ... strike that....
At the point when the problem was being demonstrated, Tim had
not been told about Atlas being rebooted. (Yes, it HAD been all
the way down.) But when he was finally told, Atlas only had about
20 minutes to go before it was all the way up. Tim thinks he was
told this inforamtion within 30 minutes of observing the user
problem.
4. On that point, what *is* the degree of repeatability?
It happened 2x in a row for Tim, and two different users said
it happened every time they tried it. He had them try it on
TAILGUNNER, the DECstation in 11-115. They had previously been
trying it on SOME public workstation(s). (Sorry, no further
info.)
5. Can the same user produce the effect on any workstation or class of
workstations while the fileserver is down? Or is the effect spurious?
Yes, see above.
This is the evidence from ONE consultant. I'll work on finding out who
else has observed this, and get them to give you the inforamtion you need
as well.
--Beth
