\section{Lost Contact With File Server}

At some point in your dealings with Athena, you will almost certainly
encounter a message saying something like:

{\tt afs: Lost contact with file server 18.70.0.210 in cell athena.mit.edu}

(The number may be different.) This is disconcerting if you've never
seen it before. But DON'T PANIC.

\subsection{What This Means To You}

(The number may be different.) This means that your workstation has lost
contact with a {\em file~server} that your workstation is using. You may
be unable to use some programs (depending on what file server you've
lost, this can be anything from your background to {\em idraw} to {\em
ls}); you may be unable to save or retrieve your files. If you've gotten
this message, can't save your files, and only want to know how to save
and/or print your files, you can skip to the end. If you want to know
why it's happening and what you can do, read the next section first.

\subsection{What Did All That Mean?}

Files in the Athena system are kept in {\em lockers} on {\em file
servers}. A locker is basically a chunk of space on a disk somewhere,
generally in building E40 or W20 or 11; a file server is a workstation,
exactly like the workstations you'll log in on, but with rather more
disk space. If a file server goes down -- for maintainance, or there's a
power outage, or terrorists take over building E40 and refuse to turn
the machines on until President Vest sings the theme to Gilligan's
Island -- you will not be able to access the files on that file server
until it comes back up. Shortly after a servers comes back up you will
see a message saying this:

{\tt afs: file server 18.70.0.210 in cell athena.mit.edu is back up}


If you have attached a locker (the sipb locker, for instance, or the
graphics locker) and the file server that contains that locker goes
down, you will get a ``lost contact'' message in your console window.
Some lockers are spread out over more than one file server so that if
one goes down you will still be able to access the information. For more
information about this, you can look in the {\em AFS} section of {\em
Inessential Athena}.


\subsection{So What Did I Lose?}

The first thing you want to do will probably be to find out which locker
you've just lost. Suppose you have just received the message:

{\tt afs: Lost contact with file server 18.70.0.210 in cell
athena.mit.edu}

The number 18.70.0.210 is the internet address of the machine; the
command {\tt hostinfo} will tell you what the name of the machine is. So
typing {\tt hostinfo 18.70.0.210} will give you this:

\indent \indent Desired host:   18.70.0.210\newline
\indent \indent Official name:  ROSEBUD.MIT.EDU\newline
\indent \indent Host address:   18.70.0.210\newline


This tells you that the name of the machine is {\em rosebud}. If you
want to know whether a particular locker is on that server, you can use
the command {\tt lookup} (in the consult locker). Thus {\tt lookup
sorokin} will return:

Filesystem sorokin (AFS) is located on:

\indent \indent ROSEMARY.MIT.EDU (read-write)\newline

which says that sorokin's homedir is still accessible. {\tt lookup sipb}
will return:

Filesystem  (AFS) is located on:

\indent \indent ROSEBUD.MIT.EDU (read-write)\newline
\indent \indent ROSEBUD.MIT.EDU (read-only)\newline
\indent \indent RONALD-ANN.MIT.EDU (read-only, alternate)\newline

This tells you that sipb is located on rosebud. Handily, there is
another file server containing the same data (ronald-ann), and so you
probably shouldn't be worried. However, you may want to look at {\em
all} the filesystems you have attached. The command for this (also in
the consult locker) is {\tt wherefs}. It will give you something like
this:

\indent \indent sorokin      : AFS filesystem on host ROSEMARY.MIT.EDU\newline
\indent \indent sipb         : AFS filesystem on host ROSEBUD.MIT.EDU\newline
\indent \indent gnu          : AFS filesystem on host ARTEMIS.MIT.EDU\newline
\indent \indent consult      : AFS filesystem on host CHICKADEE.MIT.EDU\newline
\indent \indent graphics     : AFS filesystem on host KITE.MIT.EDU\newline

There may be lockers listed that you don't recognize. The reason for
this is that when you attach a locker, it is connected to the
workstation. Therefore, a locker attached by the previous user may show
up in your listing. You can ignore it.

\subsection{So What Should I Worry About?}

First of all, servers will often come back up fairly quickly. If you
have the time and something to do, it may be worthwhile to just wait.

If the locker on the lost server is not one you're using, or is
replicated elsewhere, you can pretty much ignore the message.

If you need a program on that server, you can try to find out how long
the server will be down or whether that program exists anywhere else.
(Try asking on instance {\em help}; see the section on zephyr for
details on instances.) There may be nothing you can do but go home.

\subsection{But I Can't Save My Files!}

First of all, don't assume that not being able to save your files means
that you've lost your fileserver. Make sure that this {\em is} the
problem. 

If it is the case, you can still save your files. You just need to save
them somewhere else. In emacs, the command {\sl C-x C-w} will present
you with a pathname, and will ask you to type in the name of the file
you want to save to. Delete all of the pathname that it gives you, and
type {\tt /usr/tmp/}{\em filename}.  In EZ, simply choose the {\em Save
As} option. You have now saved your file in a temporary file on the hard
drive of your workstation. This file may hang around for up to three
days. You can mail it to yourself, with the command

{\tt /bin/mail {\em username} < {\em filename}}

{\bf DON'T INCORPORATE IT NOW. WAIT UNTIL YOU HAVE YOUR HOME DIRECTORY
BACK.}

Now, disaster averted, you can wait for the fileserver to come back up,
or you can come back later. While the server is down, you will not be
able to access any of your files.