setup.txt 6.25 KB
Newer Older
##### Setting up the Utah Network Testbed software on a boss node
##### Tested on FreeBSD 4.9

6 7 8 9 10 11 12
##### Important notes

In order to be able to help you debug any problems you run into or answer
certain questions, we'll need have accounts, preferably with root access if
allowed by your institution's AUP, on your boss and ops nodes, and will need to
be able to access the webserver on boss.

Jay Lepreau's avatar
Jay Lepreau committed
13 14 15
##### Step -1 - Set up "ops"

Follow the instructions in the setup-ops.txt file before the ones in this file!

##### Step 0 - OS installation and setup

Install FreeBSD on the machine you'll be using for your boss node, using the
20 21 22 23 24 25 26 27 28
standard FreeBSD installation process.  When asked by the installer, it's best
to choose the 'Developer' distribution set - this gets you full sources. The
'X-Developer' distribution set would be fine too, if you want to be able to run
X clients from the boss node.  When it asks if you want to install the ports
collection, answer yes.  You don't need to worry about which packages to
install (of course, grab your favorite editors, etc.) - the ones our software
needs will be installed automatically later.  You, will, however, have to make
sure that you create a partition large enough to hold /usr/testbed - in
addition to the testbed software, this is where many disk images will get
29 30 31 32
stored. The /var partition will need to be large enough to hold the database -
100MB extra for the database should be sufficient. Also, since we'll be building
and installing a lot of software from the ports tree, you'll want to make sure
that /usr is at least 2 GB.

34 35
Make sure that you have the network correctly configured, and make sure that it
has a hostname that resolves. The silly mysql-server port requires this.

37 38 39 40
If you want, you can go ahead and create an account for yourself on boss. For
now, just stick the home directory somewhere local, and move it to /users/ once
you've got it mounted from ops (the boss-install script will set this up). In
general, it's probably simpler to just use 'root' for now.

##### Step 1 - Create a defs file

44 45 46
The defs file will describe some of your setup, such as the hostnames of your
boss and ops nodes, and email addresses that certain types of mail will be sent

48 49
Use the 'defs-example' file in the root of our source distribution as a
template. It contains comments explaining the important variables to set.

##### Step 2 -  Unpacking and running configure

This works the same as it did on ops:
cd ~/tbobj
~/testbed/configure --with-TBDEFS=/users/ricci/testbed/defs-ricci

##### Step 3 - Running the boss installation script

Jay Lepreau's avatar
Jay Lepreau committed
Again, this works the same as it did on ops, except that you run
install/boss-install in the object tree, instead of ops-install.

Part way through, this script will bail out and prompt you to install some
63 64
ports. This can take a long time (hours), and you want to be able to see what's
going on, right?

66 67
So, just cd to /usr/ports/misc/emulab-boss/ and run (as root) a 'make install'.
When you're done, re-run the boss-install script.

69 70
Like the ops-install script, boss-install sets up paswordless sudo for anyone
in the wheel group.

72 73 74 75 76 77
There is one bootstrapping problem we have that needs to be worked around - we
put fully-qualified names for the ops/users node into /etc/fstab on boss. But,
if you're running the nameserver for this domain on boss, those names won't be
resolvable yet. Since we don't yet have a way to auto-generate DNS
configuration files, the suggested work-around is to add addresses for the
FSNODE and USERNODE that you specified in your defs file (which may be the same
thing), to /etc/hosts on boss. Remember to remove them once you really have
79 80
DNS set up.

##### Step 4 - Installing from source.

83 84
In your object directory, do a 'gmake && gmake boss-install'. Then, as root, do
a 'gmake post-install'. The post-install target needs to run as root, so that
it can make certain scripts setuid, etc.

##### Step 5 - Setting up root ssh from boss to ops

89 90 91 92
The boss node needs to be able to ssh in, as root, to the ops node. To set this
up, copy root's public identity from boss (created by the boss-install script)
to ops's authorized_keys file:
scp /root/.ssh/ ops:/root/.ssh/authorized_keys

##### Step 6 - Filling the database

96 97
See the file setup-db.txt in this directory for instructions on getting the
proper information about your site and nodes into the database.
Leigh B. Stoller's avatar
Leigh B. Stoller committed

##### Step 7 - Other miscellaneous things to set up

101 102
There are a few things we haven't been able to completely automate just yet,
though we hope to soon. 

104 105 106
DNS configuration - boss-install enables bind on your boss node, but does not
set it up as a server for your domain. We don't have documentation for creating
the zone files, etc yet. Best bet right now is to ask Utah for a copy of
107 108 109 110 111 112 113
theirs. You need to include boss and ops in your domain, but NOT the nodes -
these will be added for you automatically.  Once done, add as the
first nameserver in /etc/resolv.conf . Also, add boss as the first nameserver
on ops, too. If you're running under the domain, too, make sure to
get Utah to delegate the subdomain to you. If you have any hardware (eg.
switches or power controllers) on a private subnet, add them to /etc/hosts,
with the same name they've been give in the database.

115 116 117 118
hosts file - It's a good idea to put ops' name/IP address in /etc/hosts - this
helps out NFS mounts, which are typically done before the nameserver is started,
and is generally helpful if things go wrong with the nameserver.

119 120
SSL certificates - Our apache config file expects to find SSL certificates in:
/usr/local/etc/apache/ssl.crt/www.<sitename>.crt and
122 123 124 125
(where <sitename> is OURDOMAIN from the configure defs file.) 
Generate a passwordless certificate (up to you if you want to get a 'real' one
from Verisign, etc., or sign your own), and place the files from it in the
above locations.

127 128 129 130 131
tftpboot - There are a few bootloaders, mini-kernels, and MFSes that are used
to boot, reload, etc. testbed machines, which live in /tftpboot . For the time
being, these are not distributed with our source, and require some site
customizations, so ask Utah for the boot loaders, etc.

disk images - You'll also, of course, need disk images to go on your nodes.
133 134
Right now, we have no automatic way of generating these, so you'll have to ask
Utah for some.