changes.4-82

.TL
Surviving with 4.1a bsd
.AU
Bill Joy
Sam Leffler
.AI
Computer Systems Research Group
Department of EECS
U.C. Berkeley
Berkeley, Ca.  94720
.SH
Introduction
.PP
.de IR
\fI\\$1\fR\\$2\\fR
..
This short paper describes what ``4.1a bsd'' is,
and what you need to know to make best use of this
release of the system if it is running on your machine.
.SH
What is 4.1a?
.PP
4.1a is an intermediate version of the VAX system distributed
by our group.  It is an
.I experimental
version of the system whose most noticeable improvement over
the 4.1 system is its support for local networking.
It also includes a few other features which you may find useful,
such as ``symbolic links'' and an improved group scheme.
.PP
4.1a is not a full distribution of the VAX system.
It is being used at a
.I very
limited number of sites, and is
.B not
being made available to the general VAX/UNIX user community.
.PP
We currently plan to have two more intermediate versions of the VAX
system before the next full release.  The next intermediate
version 4.1b will tenatively include a higher-performance file
system, a more fully integrated scheme for inter-process
communication, and page-sharing virtual memory facilities.
These facilities are described in a draft new system manual
which is available from our group.
.PP
The rest of this document consists of a number of sections
describing new commands and changes to existing facilities
in the new system.
.NH 1
User commands for networking
.PP
The new system supports the DARPA standard TCP, IP and UDP communications
protocols.  These are used to provide for the transfer of data
between machines interconnected via local networks (such as an ETHERNET)
or a long-haul network such as the ARPANET.  Currently most
hosts on the ARPANET speak older protocols, so you are likely to find
the local networking facilities to be of the most immediate use.
.PP
Most of the networking commands provided with this 4.1a release
begin with a letter
.B r
standing for ``remote''.  These commands are all provisional,
and subject to change.
.NH 2
Status of the network
.PP
Each machine on your local network is kept informed by other machines
as to the status of those machines.  The command
.I ruptime
prints a status line for each machine on the local network.
Its output might look like:
.DS
monet % ruptime
arpa      up  1+10:22,     1 user,   load 0.94, 0.80, 0.70
calder       up  0:57,     0 users,  load 0.20, 0.75, 1.08
dali         up  2:23,     0 users,  load 0.08, 0.10, 0.09
ingres    up  1+17:26,     0 users,  load 0.32, 0.49, 0.67
kim       up  1+10:27,     0 users,  load 0.74, 0.75, 0.77
matisse    down 13:36
monet        up  7:01,     1 user,   load 0.73, 0.79, 0.78
monet %
.DE
Here ``monet % '' is my prompt.
This command shows 7 hosts on the local network.  For each
host the number of users logged in, the load average and the
amount of time the machine has been up (or down) is shown.
As it is very late (about 3 am), there are essentially no users
logged in.  In fact, the users which are logged in will be shown
by a
.DS
monet % rwho
root     arpa:ttyh9  Mar 31 23:42
root     monet:ttyp0 Mar 31 23:39
monet %
.DE
command.  Currently this shows only me (logged in a root on
the \fImonet\fP machine), and another \fIroot\fP user (logged
in on the \fIarpa\fP machine.)
.PP
Both of these commands suppress and do not count users who have
been idle more than an hour.  They both take \fB\-a\fP commands
to force these users to be printed.  Thus
.DS
monet % rwho \-a
fabry    arpa:ttyh3    Mar 30 17:59 10:51
kateveni kim:ttyh4     Mar 30 18:42  9:10
kridle   ingres:ttyh0  Mar 31 18:48  6:32
mckusick arpa:ttyj1    Mar 31 19:50  6:50
mosher   arpa:ttyh1    Mar 31 08:26  2:48
root     arpa:console  Mar 31 12:19  1:55
root     arpa:ttyh9    Mar 31 23:42
root     monet:console Mar 31 20:54  6:45
root     monet:ttyp0   Mar 31 23:39
root     monet:ttyp1   Apr  1 01:45  2:11
rwh      arpa:ttyh5    Mar 31 14:34 13:20
sam      arpa:ttyh2    Mar 31 11:36 15:50
wilensky kim:ttyib     Mar 31 11:17 10:22
wnj      monet:tty00   Mar 31 21:17  6:38
monet %
.DE
produces more, if rather pointless, output.
.NH 2
Logging in and executing commands on remote machines
.PP
It is quite easy to log in on another machine and execute commands
there.  You can do this by saying
.DS
monet % rlogin \fIothermachine\fP
.DE
If your current machine and the \fIothermachine\fP name that you
give are under common administration you may find that you are
simply logged in on the other machine with no fuss.
If not, you may get a prompt for a password.*
.FS
*If there is not
an account on that machine with your current login name, then
you will likely be prompted for a login name.  Other possibilities
are mentioned in
.IR rlogin (1x).
.FE
Here, for example, is what is printed when I log into the
\fIkim\fP machine from my current machine \fImonet\fP:
.ID
.nf
monet % rlogin kim
Last login: Wed Mar 31 20:32:25 on ttyp0
TERM = (c1004p) h19
Erase is Delete
Kill is Ctrl-U
  4:10am  up 1 day, 10:34,  3 users,  load average: 0.64, 0.67, 0.73
kim % who
kateveni ttyh4   Mar 30 18:42
wilensky ttyib   Mar 31 11:17
wnj      ttyp0   Apr  1 04:10
kim % date
Thu Apr  1 04:10:33 PST 1982
kim % hostname
ucbkim
kim % logout
Lost connection.
monet % 
.fi
.DE
On the
\fIkim\fP machine my login prompt includes the machine name,
just as it does on the \fImonet\fP machine.  I do this while having
a single \fI.cshrc\fP file by using the
\fIhostname\fP command, in a sequence of the form:
.DS
if ($?prompt) then