The following is an incomplete guide to compiling, setting up and using SKS. Hopefully this is enough to get you started, in addition there is a wiki available, where in particular https://bitbucket.org/skskeyserver/sks-keyserver/wiki/Peering should help getting a working installation.
There are a few prerequisites to building this code. You need:
- OCaml-4.0 or later. Get it from http://ocaml.org
- Berkeley DB version 4.6.* or later. You can find the appropriate versions at http://www.oracle.com/technetwork/database/berkeleydb/downloads/index.html
- GNU Make and a C compiler (e.g gcc)
Verifying the integrity of the download
Releases of SKS are signed using the SKS Keyserver Signing Key available on public keyservers with the KeyID
and has a fingerprint of
C90E F143 0B3A C0DF D00E 6EA5 4125 9773 973A 612A.
Using GnuPG, verification can be accomplished by, first, retrieving the signing key using
gpg --keyserver pool.sks-keyservers.net --recv-key 0x41259773973A612A
followed by verifying that you have the correct key
gpg --keyid-format long --fingerprint 0x41259773973A612A
pub 4096R/41259773973A612A 2012-06-27 Key fingerprint = C90E F143 0B3A C0DF D00E 6EA5 4125 9773 973A 612A
A check should also be made that the key is signed by trustworthy other keys;
gpg --list-sigs 0x41259773973A612A
and the fingerprint should be verified through other trustworthy sources.
Once you are certain that you have the correct key downloaded, you can create a local signature, in order to remember that you have verified the key.
gpg --lsign-key 0x41259773973A612A
Finally; verifying the downloaded file can be done using
gpg --keyid-format long --verify sks-x.y.z.tgz.asc
The resulting output should be similar to
gpg: Signature made Wed Jun 27 12:52:39 2012 CEST gpg: using RSA key 41259773973A612A gpg: Good signature from "SKS Keyserver Signing Key"
Compilation and Installation
Install OCaml and Berkeley DB
When installing ocaml, make sure you do both the
make worldand the
make optsteps before installing. The later makes sure you get the optimizing compilers. (do make opt.opt if you want faster compilation. You can then set the environment variables
If your vendor or porting project supplies prebuilt binaries and libraries for Berkeley DB, make sure to get the development package as you will need the correct version include files.
Makefile.local, and edit to match your installation.
make dep make all make all.bc # if you want the bytecode versions make install # puts executables in $PREFIX/bin, as defined # in Makefile.local
There are some other useful compilation targets, mostly useful for development.
creates a doc directory with ocamldoc-generated documentation of the individual modules. These are mostly useful as documentation to the source code, not a user's guide.
Creates a ps-file that shows the dependencies between different modules, and gives you a sense of the overall structure of the system. For this to work you need to have AT&T's graphviz installed, as well as python2. The python script that's used actually requires that python2 be called python2, rather than python. You can of course edit that script.
Setup and Configuration
You need to set up a directory for the SKS installation. It will contain the database files along with configuration and log files.
Configuration options can be passed in on the command-line or put in
sksconf file in the SKS directory. the
specifies the SKS directory itself, which defaults to the current
Sksconf and commandline options
The format of the sksconf file is simply a bunch of lines of the form:
# character is used for comments, and blank lines are
ignored. The keywords are just the command-line flags, minus the
The one thing you probably want no matter what is a line that says
which ensures that sks will output messages to
If you want your server to gossip with others, you will need a
membership file which tells the
sks recon who else to gossip with.
The membership file should look something like:
epidemic.cs.cornell.edu 11370 athos.rutgers.edu 11370 ...
This file should be called
membership, and should be stored in the
SKS directory. Note that in order for synchronization to work, both
hosts have to have each other in their membership lists. Send mail to
email@example.com to get other SKS administrators to add you to
their membership lists.
IMPORTANT NOTE: if you include the server itself in the membership
file, you should make sure that you also specify the
option, and that the selected hostname is exactly the same string
listed in the membership file. Otherwise, the
sks recon will try to
synchronize with itself and will deadlock.
Outgoing PKS synchronization: mailsync file
The mailsync file contains a list of email addresses of PKS keyservers. This file is important, because it ensures that keys submitted directly to an SKS keyserver are also forwarded to PKS keyservers.
IMPORTANT: don't add someone to your mailsync file without getting their permission first!
In order for outgoing email sync's to work, you need to specify a
command to actually send the email out. The default is
-oi, but you may need something different.
Incoming PKS synchronization
Incoming PKS synchronization is less critical than outgoing, since as long as some SKS server gets the new data, it will be distributed to all. Having more hosts receive the incoming PKS syncs does, however, increase the fault-tolerance of the connection between the two systems.
In order to get incoming mail working, you should pipe the appropriate incoming mail to the following command via procmail:
Here's an example procmail entry:
PATH=/path/of/sks/exectuables :0 * ^Subject: incremental | sks_add_mail sks_directory_name
You can server up a simple index page directly from the port
you're using for HKP. This is done by creating a subdirectory in
your SKS directory called
web. There, you can put an index file
supporting files with extensions .css, .es, or .js, and some image
files with extensions jpg, jpeg, png or gif. Subdirectories will
be ignored, as will filenames with anything other than
alphanumeric characters and the '.' character. This is
particularly useful if you want to run your webserver off of port
80. This can be done by using the -hkp_port command-line option.
Building up the databases
First, you need to get a keydump. If you're running a PKS server, you should be able to convince PKS to generate one for you. If you're starting from scratch, you'll need to download one from the net. You should contact the pgp keyserver list firstname.lastname@example.org
in the SKS directory, put in a subdirectory called
dumpwhich contains the keydump files from which the database is to be built.
Run sks_build.sh. That script actually runs three utilities. You might want to edit sks_build.sh if you want to trade off speed for space usage. At the current settings, you could run out of ram if you try this with less then 256 megs of RAM.
DO NOT DELETE THE
dump DIRECTORY, even after the database is
built. The original keys are not copied to the database, and so the
dump must be left in place.
Platform specific issues
On FreeBSD it appears that libdb is named differently than on some
other platforms. For that reason, you need to set the LIBDB
environment value to
-ldb46 instead of
-ldb-4.6 for other