[5652] in Hotline Meeting
Your OLC question about "other"
daemon@ATHENA.MIT.EDU (kevles@ATHENA.MIT.EDU)
Thu Jul 18 14:17:44 1991
From: kevles@ATHENA.MIT.EDU
Date: Thu, 18 Jul 91 14:17:33 -0400
To: hotline@ATHENA.MIT.EDU
Cc: gejaros@ATHENA.MIT.EDU
--------
To hotline:
One of the Woods Hole workstations has a problem (for a month now) and
the cluster manager there is away. Here's some log extracts to provide
you with clues. Can you please look into this?
Thanks.
--Beth
(The workstation is MENHADEN)
For some reason I can no longer access the sipb, rrn, or games files.
For example:
I get an aklog returned a bad exit status (6)
And when I logout I get a "login incorrect" message.
Hi,
I get the following:
aklog: Can't get information about cell athena.mit.edu.
aklog: Can't get information about cell athena.mit.edu.
This has actully been happenning for few weeks, so it may not be related to thedown-server.
After typing ls /afs, i get back the Athena prompt by itself.
Hello again.
Well...1. When I type 'fs listcells', I get the prompt back.
2. When I type cd /usr/vice/etc
more CellServDB
I get the following message:
Fixed up unaligned data access for pid 1322 (more) at pc 0x400d80
3. I rebooted and still get:
18:55 sipb: aklog returned a bad exit status (6)
18:55 sipb: /afs/athena.mit.edu/contrib/sipb does not exist
--- Comment by consultant dkk@TISIPHONE.MIT.EDU [0].
[Thu 04-Jul-91 12:45pm]
The problem isn't that Menhaden is far away (in Woods Hole) or that it
has a slow network link to campus. (The network link in anything but
slow - It's as good as some on campus.) I checked the AFS access of
Lobster, on of the WHOI (Woods Hole Oceanographic Institute) Athena
servers, and found its AFS performance was just fine. The problem
is not WHOI, just Menhaden.
Any of a number of causes could be responsible for the current AFS
problem. Menhaden appears to have been privatized (it accepts rsh.ucb
connections) and may have been mis-configured. More likely, the
CellServDB was just lost, and needs to be copied in via NFS or rsh.
About the unaligned access error... Mark Eichin tells me there is
a known bug in "more" that sometimes clashes with the decmips memory
alignment constraints. Some other program (such as "cat" or "diff")
should be used on CellServDB.
*** Mail from consultant dkk@TISIPHONE.MIT.EDU [0].
[Thu 04-Jul-91 12:58pm]
To: Gary E Jaroslow <gejaros@Athena.mit.edu>
cc:
Subject: Your OLC question about "other"
--------
Are you still unable to access /afs from Menhaden? I came across your
OLC question, and looked into it some. I think you quite likely have
a bad copy of /usr/vice/etc/CellServDB . To check the file, type the
following commands at the "athena%" prompt:
attach bitbucket
diff /mit/bitbucket/menhaden/CellServDB /usr/vice/etc/CellServDB
If the "diff" command gives you any output, Menhaden's AFS
configuration file (CellServDB) is out of date or has been damaged.
The correct CellServDB (the one in /mit/bitbucket/menhaden) has about
140 lines, so if you get screensful of "diff" output, you local file
most probably needs to be replaced, then the workstation rebooted.
Warning: If you or a WHOI system administrator plan to replace the
file yourselves, note that /usr/vice/etc/CellServDB is just a symbolic
link to CellServDB.public, and that CellServDB.public is that file
that should be replaced (so that the file will be updated properly
when the master copy is changed).
If you have any doubt at all about proceeding with such a fix, please
follow-up in OLC and/or have a systems administrator contact Athena
Systems Support. It's much better if you (or a sysadmin) are "talked"
through a fix, than if you attempt it alone and make things worse.
I hope we can get this fixed soon...
David Krikorian
Athena alumnus
SIPB member
Hi,
Regarding a messdage sent to me from Carla...The Woods Hole Athena
consulata nt, Julie Allen, has been away. Now I'm leaving for a month.
I am passing the i nformation on to another user to work it out in the
meantime. Thanks, Gary
=========================================================================
