This is a script and a set of template files which customises a generic 4.3BSD SimH disk image so that it acts as a uucp node and connects to other uucp nodes via TCP links.
You will need the bsdtar program installed so that tarballs compatible with
4.3BSD can be installed. On Ubuntu, sudo apt-get install bsdtar. Can someone
add instructions for other systems? The source for bsdtar is at
http://www.libarchive.org/
You will also need a recent version of SimH: 4.0; version 3.9 doesn't work. You can download the SimH Github repository at https://github.com/simh/simh. In your local copy, build a vax780 SimH binary and copy the resulting binary somewhere useful:
make vax780
sudo cp BIN/vax780 /usr/local/bin
Download this Github repository. In this repository, compile the mktape program:
cc -o mktape mktape.c
You will see the generic 4.3BSD SimH image, rq.dsk.gz. The buildimg script
builds a tarball for each uucp system with the specific changes for that
system.
As an example, look at the site_generate script:
# An example of three uucp sites connected in series
#
# site5 ----- site6 ----- site7
#
./buildimg site5:5000 site6:localhost:6000
./buildimg site6:6000 site5:localhost:5000 site7:localhost:7000
./buildimg site7:7000 site6:localhost:6000
site5 allows you to telnet in on TCP port 5000. Ditto site6 and TCP port 6000, and site7 and TCP port 7000.
site5 has a (simulated) hard-wired serial connection to site6 through TCP port 6000. The syntax allows site6 to be on a remote computer, e.g. available through www.somewhere.com:6000.
site6 has a (simulated) hard-wired serial connection to site5 and site7. Finally site7 has a (simulated) hard-wired serial connection to site6. This means that each site can initiate a connection to another site; they don't have to wait for the other site to dial in.
For each site, there are two files generated:
siteX.iniis the SimH config file to run this systemsiteX.tapis a tarball in tap format with the customisations
We also need to copy the generic disk image rq.dsk.gz to be the disk
for each site:
zcat rq.dsk.gz | dd conv=sparse > site5.dsk
zcat rq.dsk.gz | dd conv=sparse > site6.dsk
zcat rq.dsk.gz | dd conv=sparse > site7.dsk
I prefer to use dd conv=sparse because a lot of the disk image is empty
and this saves disk space for these files.
Once you have a disk, a config file and the customisation tape, here is how you customise the system. Boot up the system:
vax780 site5.ini
At the login: prompt, login as root with no password. At the shell prompt,
read in and unpack the tarball, and run a script which sets appropriate
file permissions:
myname# tar xf /dev/rmt12
tar: blocksize = 1
myname# ./mkdirs
29 password entries, maximum length 93
Type ctrl-D at the # prompt to restart things
erase ^?, kill ^U, intr ^C
#
Type a ctrl-D to restart in multi-user mode and login as root. Your system now has a hostname:
login: root
Last login: Wed Mar 7 10:52:10 on console
You have mail.
Don't login as root, use su
site5#
Repeat the process for the other sites, e.g. site6 and site7. SET ROOT PASSWORDS NOW!! It is a good idea to reboot your simulated system, as this will pick up the new kernel from the tarball.
On site5 you can send e-mail to site6!root and site6!site7!root.
Watch out for escaping the bang characters as the shell is csh, e.g.
echo Hello there | mail site6\!site7\!root
You should be able to work out the bangpaths to send e-mails on the other systems.
Your mail is now sitting on your system waiting to be delivered over uucp. Before you try to do a uucp connection, you can check if you have a working serial link with a remote uucp site. In your simulated 4.3BSD system, edit the dialer line in /etc/remote to say:
dialer:dv=/dev/tty00:br#9600:
Now try:
# tip dialer
which should connect out over /dev/tty00 to the remote uucp site via the
TCP connection. Hit Return a few times to see if there is any response. On your
host system, do netstat -a | grep ESTAB and see if there is a TCP connection
to the remote system. To get out of tip, type in the two characters ~.
Here is how to perform a manual uucp connection. On site5, to call site6:
# /usr/lib/uucp/uucico -r1 -ssite6 -x7
root site6 (3/9-06:27-166) DEBUG (Local Enabled)
finds (site6) called
getto: call no. tty00 for sys site6
Using DIR to call
Opening /dev/tty00
login called
wanted """"
got: that
. . .
Password:got: that
send "uucp"
root site6 (3/9-06:27-166) SUCCEEDED (call to site6 )
imsg looking for SYNC<
\20>
imsg input<Shere=site6\0>got 11 characters
omsg <Ssite5 -Q0 -x7>
imsg looking for SYNC<
. . .
Proto started g
protocol g
root site6 (3/9-06:27-166) OK (startup tty00 9600 baud)
*** TOP *** - role=MASTER
daemon site6 (3/9-06:27-166) REQUEST (S D.site5B00D2 D.site5S00D2 daemon)
. . .
PROCESS: msg - SY
SNDFILE:
send 0221
send 0231
send 0241
rec h->cntl 042
state - 010
. . .
daemon site6 (3/9-06:27-166) OK (conversation complete)
send OO 0,omsg <OOOOOO>
imsg looking for SYNC<\0\0\20>
imsg input< \5*%\3\20>
imsg input< "*\10 \20>
imsg input< "*\10 \20>
imsg input<OOOOOO\0>got 6 characters
site5#
The -x7 flag turns on debugging; it's not normally used in production.
Now go to site6 and run a similar command to forward the e-mail to
site7:
# /usr/lib/uucp/uucico -r1 -ssite7 -x7
On site7, you can now read your e-mail:
site7# mail
Mail version 2.18 5/19/83. Type ? for help.
"/usr/spool/mail/root": 3 messages 3 new
>N 1 MAILER-DAEMON Sat Jul 9 23:21 35/1191 "Returned mail: Host unknown"
N 2 MAILER-DAEMON Sat Jul 9 23:22 33/986 "Returned mail: Host unknown"
N 3 site5!root Wed Mar 7 11:02 14/422
& 3
Message 3:
From site6!site5!root Wed Mar 7 11:02:14 1984
Received: by site7.ARPA (4.12/4.7)
id AA00173; Wed, 7 Mar 84 11:02:14 pst
Received: by site6.ARPA (4.12/4.7)
id AA00168; Wed, 7 Mar 84 10:58:33 pst
Received: by site5.ARPA (4.12/4.7)
id AA00170; Wed, 7 Mar 84 10:55:27 pst
Date: Wed, 7 Mar 84 10:55:27 pst
From: site6!site5!root (Charlie Root)
Message-Id: <[email protected]>
To: site6!site7!root
Status: R
Hello there
&
You can edit /usr/lib/crontab to have entries that run uucico for each site
that you connection to. Here is an example line that connects to site6
every minute:
* * * * * uucp /usr/lib/uucp/uucico -r1 -ssite6
The system has to dedicate a /dev/tty0x device for each outbound uucp connection. If you have this situation, read through the tcpdial documentation for a solution.
The mail system uses pathalias and smail, as well as sendmail. These allow you to send e-mail to a uucp site several hops away like this:
$ echo hello | mail [email protected]
$ echo hello | mail site!user
$ echo hello | mail long!bang!path!to!site!user
smail needs to derive a set of best paths to all the remote sites on the
global uucp network. To do this, copy the uucp.map file into your
system as the file /usr/lib/uucp/uucp.map. Then run the
/usr/lib/uucp/mkpaths script to create the file /usr/lib/uucp/paths.
This holds the best path to deliver mail to each site.
If you see very high numbers, that's a pathalias bug. Edit the file and set the number down to e.g. 10000.
The buildimg script will configure your system to be ready to run C News, but you need to set up some cron jobs to actually make it happen. Read through the C News documentation to see how to do this.
If you want to install other applications, you'll have to find them, compile them and install them. There is an archive of applications already configured at http://www.tuhs.org/Uucp/43BSD_Apps. You might also want to look at an archive of the http://ftp.acc.umu.se/mirror/archive/ftp.sunet.se/pub/doc/usenet/ftp.uu.net/comp.sources.unix/ Usenet postings.
The tty lines are exposed to the Internet through the bound TCP port, so you may want to implement some firewall rules to only allow connections from specific IP addresses.
The tty lines are set as insecure in the 4.3BSD /etc/ttys, so they won't
allow root logins. You can only login as root on the console, i.e. the place
where you ran vax780 system.ini. You should probably do this in a tmux
or screen session, so that you can get back to the 4.3BSD console easily.
It is a good idea to add a non-root user
so that you can telnet in on the TCP port: only do this on localhost, as
the telnet session is not encrypted. If you add this non-root user to the
group wheel (in /etc/group), then you can su and become root.
Make sure that you are running the kernel from the buildimg-created tarball: you will see munnari ... 2017 as the kernel boots. The tty lines in the new kernel are set to kill off running shells when an incoming telnet session is closed or disconnects unexpectedly.
If you are running a SimH site which accepts connections from non-SimH uucp sites (or vice versa), then there can be a problem because SimH uses the Telnet protocol on the TCP port to, for example, know when to echo/not echo text (think: passwords). This can cause uucp protocol problems, as the Telnet can interpret the incoming uucp data, which screws up the uucp protocol.
If you have this situation, you can set up an incoming TCP port with Telnet disabled. You will still want some of your simulated tty lines doing Telnet, so that you can log into your simulated 4.3BSD system and have your password hidden. Here is an example SimH configuration file with Telnet disabled.
# Set up eight DZ serial ports in 8-bit mode.
# Connect to a remote uucp site on 127.0.0.1:6000
# Listen on TCP port 5001 with Telnet disabled
# All other DZ lines will listen on port 5000 with Telnet enabled
set dz lines=8
set dz 8b
attach dz -a -m line=0,Connect=127.0.0.1:6000
attach dz -a -m line=1,5001;notelnet
attach dz -a -m 5000
At present, none!
If you are interested in joining this simulation of the 1980s uucp network, then send some e-mail to Warren Toomey. Indicate what historical uucp site(s) you want to run and which other sites you want to connect to.
If you don't know what historical site to choose, have a look at the maps here: http://www.redace.org/html/logical_usenet_map_1984.html
If you have a specific historical site in mind, do a search on the site's name, e.g. "#N decvax", and include the double quotes; add the uucp keyword as well if you like. Look for the lines after the "#N decvax" without hash characters: they show you the connectivity that this site used to have. An example for decvax:
#N decvax, decvax.dec.com
#S Vax 8300; Ultrix 32 V2.2
#O Digital Equipment Corp.
#C Larry Palmer
#E decvax!lp
#T (603)8848385
#P MK2-1/H10 Continental Blvd, Merrimack, NH 03054
#L 42 49 N / 71 31 W
#R
#W
#U bellcore cca mandrill dartvax decwrl genrad gsg harpo ichaya
#U linus mcnc savax tektronix ucbvax vortex yale
#R decvax has moved to a new location - an update is expected soon... -jj
#
# local calls
decvax bu-tyng(DIRECT), ichaya(DIRECT), savax(DIRECT), sequitor(DIRECT),
shaman(DIRECT), sii(DIRECT), skippy(DIRECT), stellar(DIRECT)
# Internet
decvax decwrl(DEDICATED), ucbvax(DEDICATED)
#
# frequent calls
decvax bellcore(HOURLY), cca(HOURLY), cg-atla(HOURLY), mandrill(HOURLY),
deceds(HOURLY), decuac(HOURLY), genrad(HOURLY), granite(HOURLY),
hanauma(HOURLY), ihnp4(HOURLY), lafite(HOURLY),
linus(HOURLY), mcnc(HOURLY), mit-athena(HOURLY), netrix(HOURLY),
tektronix(HOURLY), vortex(HOURLY), yale(HOURLY)
#
# non-prime time calls
decvax allegra(EVENING), astrovax(EVENING), bobo(EVENING), brunix(EVENING),
bunker(EVENING), cbosg(EVENING), cbosgd(EVENING), cincy(EVENING),
cornell(EVENING), cray(EVENING), duke(EVENING), emory(EVENING),
encore(EVENING), esquire(EVENING), farance(EVENING), freeport(EVENING),
gatech(EVENING), hcr(EVENING), hermix(EVENING), hjuxa(EVENING),
hplabs(EVENING), idis(EVENING), ima(EVENING),
masscomp(EVENING), microsoft(EVENING), mit-vax(EVENING),
mtxinu(EVENING), netword(EVENING), philabs(EVENING), pur-ee(EVENING),
purdue(EVENING), randvax(EVENING), research(EVENING),
rochester(EVENING), sdcsvax(EVENING), sickkids(EVENING), std(EVENING),
stolaf(EVENING), sun(EVENING), trwrb(EVENING), ucf-cs(EVENING),
ulysses(EVENING), usenix(EVENING), utah-cs(EVENING), utzoo(EVENING),
uw-beaver(EVENING), vax135(EVENING), watmath(EVENING)
#
# very long distance calls
decvax mulga(DAILY)
#
# incoming calls
decvax apollo(POLLED), attunix(POLLED), bbn(POLLED), bhjat(POLLED),
brl-smoke(POLLED), chaos(POLLED), cvbnet(POLLED), dartvax(POLLED),
elrond(POLLED), epiwrl(POLLED), flkvax(POLLED),
frog(POLLED), gsg(POLLED), harpo(POLLED), howtek(POLLED),
humus(POLLED), jaxlab(POLLED), ll-xn(POLLED), mindcrft(POLLED),
mkunix(POLLED), necntc(POLLED), nicmad(POLLED), noao(POLLED),
nrl-css(POLLED), plato(POLLED), raster(POLLED), stcvax(POLLED),
sunybcs(POLLED), tifsie(POLLED), unh(POLLED), virgin(POLLED),
wang(POLLED)
For central sites (like decvax) that had a lot of connectivity, you will be expected to run them continuously. For edge sites which only dial in to one other site to exchange news and e-mail, you can run them whenever you want.