This is a collection of notes on how to complete tasks I've found myself
commonly performing while using 9front.
Most of what is here is straight from various parts of the 9front FQA, just
ordered differently. If not from the FQA, I'll provide the source under the
header for the section.
As I'm in the process of learning Plan 9/9front, feel free to contact me
with suggestions or corrections.
I'm overly verbose at times because I'm extremely forgetful.
____
/_ | Building & installing the amd64 kernel
| | ______________________________________________________________________
| |
|___|
% cd /
% rc /sys/lib/rootstub
% cd /sys/src
% objtype=amd64 mk install
% cd /sys/src/9/pc64
% mk install
Now the amd64 kernel has been built. First, install the new bootloader to
the 9fat partition:
Following these steps sets up a machine with 9front installed as a cpu AND
auth server. I don't cover fileserver configuration since I've just been logging
in as the default user, but it is covered in FQA 7.3
For those of us coming from the *nix world... setting up a machine running
9front as a cpu & auth server will allow you to remote into it using drawterm
(covered later) and have access to the rio GUI. Similar to X11 forwarding using
ssh -X. I don't currently know the nitty-gritty of how it works, but remote
access was something I wanted to set up right away.
First, add the line service=cpu to /n/9fat/plan9.ini. This defines the service
variable during the boot process so the cpu kernel is booted.
Reboot.
Upon rebooting, you'll get something along the lines of the following:
bad nvram key
bad authentication id
bad authentication domain # You may not see these errors.
authid: <glenda>
authdom: <9front>
secstore key:
password: [glenda’s password]
Enter `glenda` for authid, `9front` (or anything you want) for authdom, hit
return for secstore key, and enter a password.
Now you'll be in an rc shell only, no rio GUI. You can run rio -i riostart if
you prefer to make the rest of the changes using acme or sam, but I tend to just
use ed to make the edits via the command line.
If you installed the cwfs filesystem instead of hjfs, reboot with fshalt -r.
When prompted for bootargs during boot, enter `local!/dev/sdXX/fscache -c`,
where XX corresponds to whatever your drive identifier is. Then type `noauth`
twice, followed by `end`:
config: noauth
auth is now disabled
config: noauth
auth is now enabled
config: end
Finally, run `9fs 9fat` and edit /n/9fat/plan9.ini. Change the bootargs variable
to one of the following.
* for cwfs -> bootargs=local!/dev/sdXX/fscache -a tcp!*!564
* for hjfs -> bootargs=local!/dev/sdXX/fs -m 702 -A -a tcp!*!564
* 702 will probably be a different value on your install, just don't
change the value from what it was by default.
Reboot. The cpu server is now listening for connections.
We need to edit the file /lib/ndb/local, but this was a bit tricky for me at
first. The contents of this file on a fresh install of 9front-5966.df2c352ae695:
#
# files comprising the data,base, use as many as you like, see ndb(6)
#
database=
file=/net/ndb
file=/lib/ndb/local
file=/lib/ndb/common
#
# entries defining the dns root. these will be overridden by any
# authentic info obtained from the root.
#
dom=
ns=A.ROOT-SERVERS.NET
ns=B.ROOT-SERVERS.NET
ns=C.ROOT-SERVERS.NET
ns=D.ROOT-SERVERS.NET
ns=E.ROOT-SERVERS.NET
ns=F.ROOT-SERVERS.NET
ns=G.ROOT-SERVERS.NET
ns=H.ROOT-SERVERS.NET
ns=I.ROOT-SERVERS.NET
ns=J.ROOT-SERVERS.NET
ns=K.ROOT-SERVERS.NET
ns=L.ROOT-SERVERS.NET
ns=M.ROOT-SERVERS.NET
#
# because the public demands the name localsource
#
ip=127.0.0.1 sys=localhost dom=localhost
# example: adjust to fit your network
#auth=cirno authdom=9front
#ipnet=9front ip=192.168.0.0 ipmask=255.255.255.0
# ipgw=192.168.0.1
# dns=192.168.0.1
# auth=cirno
# dnsdom=9front
# cpu=cirno
# smtp=cirno
#
#ip=192.168.0.99 sys=cirno dom=cirno.9front ether=112233445566
sys=cirno ether=000c29e16a29
Before making the changes, run `ndb/query sys SYSNAME` to get some info about
your network. It will output something like:
Use this information to edit /lib/ndb/local so it looks similar to:
#
# files comprising the data,base, use as many as you like, see ndb(6)
#
database=
file=/net/ndb
file=/lib/ndb/local
file=/lib/ndb/common
#
# entries defining the dns root. these will be overridden by any
# authentic info obtained from the root.
#
dom=
ns=A.ROOT-SERVERS.NET
ns=B.ROOT-SERVERS.NET
ns=C.ROOT-SERVERS.NET
ns=D.ROOT-SERVERS.NET
ns=E.ROOT-SERVERS.NET
ns=F.ROOT-SERVERS.NET
ns=G.ROOT-SERVERS.NET
ns=H.ROOT-SERVERS.NET
ns=I.ROOT-SERVERS.NET
ns=J.ROOT-SERVERS.NET
ns=K.ROOT-SERVERS.NET
ns=L.ROOT-SERVERS.NET
ns=M.ROOT-SERVERS.NET
#
# because the public demands the name localsource
#
ip=127.0.0.1 sys=localhost dom=localhost
# example: adjust to fit your network
auth=cirno authdom=9front
ipnet=9front ip=192.168.0.0 ipmask=255.255.255.0
ipgw=192.168.0.1
dns=192.168.0.1
auth=cirno
dnsdom=9front
cpu=cirno
fs=cirno
# smtp=cirno
#
#ip=192.168.0.99 sys=cirno dom=cirno.9front ether=112233445566
sys=cirno ether=000c29e16a29
NOTE: The indentation of the lines under ipnet=9front ip=192.168.0.0
ipmask=255.255.255.0 DOES matter.
A more detailed explanation:
* auth and cpu must be set to the system name.
* authdom can be anything, but has to match the authdom we entered at the
boot prompt earlier.
* ipnet should match authdom.
* ipmask, ipgw, and dns should be changed to values from the output of
`ndb/query sys <sysname>`.
* You'll notice ip does NOT match the output from ndb/query. In
/lib/ndb/local, ip must be set to the subnet ID of the network. This just
means the last digit is 0. For example, if the output of ip from
`ndb/query sys <sysname>` was 156.43.9.78, ip in /lib/ndb/local
should be set to 156.43.9.0.
Reboot.
Start keyfs by running `auth/keyfs`.
Run `auth/changeuser glenda` to get the following:
Password: # type password here, will not echo
Confirm password: # confirm password here, will not echo
assign Inferno/POP secret? (y/n) n
Expiration date (YYYYMMDD or never)[return = never]:
2 keys read
Post id:
User’s full name:
Department #:
User’s email address:
Sponsor’s email address:
user username installed for Plan 9
NOTE: Password MUST match what you entered for glenda earlier.
After "assign Inferno/POP secret? (y/n) n" you can just hit enter for everything
else.
If everything is set up correctly, running `netaudit` should output something
similar to:
checking this host's tuple:
ip=192.168.1.23 looks ok
no dom= entry
ether=000c29e16a29 looks ok
checking the network tuple:
we are in ipnet=9front
ipgw=192.168.1.1 looks ok
dns=192.168.1.1 look ok
auth=cirno looks ok
checking auth server configuration:
we are the auth server
auth/keyfs is running
someone is listening on port 567
run auth/debug to test the auth server
run auth/asaudit to verify auth server configuration
You'll probably want to be able to access websites and have rio running when
you connect via drawterm. While you can add whatever you like to your profile,
I recommend adding the following to /usr/glenda/lib/profile:
webfs
bind -b '#S' /dev
rio -i riostart
Add those right below fn cpu% { $* } and you'll be able to interact with
URLs, mount 9fat or other filesystems via 9fs, and rio will start automatically
on connect.
Download and install 9front's fork of drawterm on the machine you want to
connect to the 9front server from. As stated on the website, this version of
drawterm is needed to connect to a 9front server due to some security protocol
changes. Russ Cox's version won't work... that tripped me up initially.
Run `drawterm -h <server_ip> -u glenda`. A rio window with a prompt should
appear. Hit return at the auth prompt, then type the password for glenda.
We're in!
