The Configuration File
The first step in configuring a
proftpd daemon is knowing where
the configuration file, usually named
proftpd.conf, is located.
The default location for this file is
/usr/local/etc/proftpd.conf, depending on your installation.
When starting the daemon, the exact path to the configuration file to be
used can always be specified using the
-c command-line option.
The idea behind
proftpd's handling of the configuration file
is that a blank file can be used, and the daemon will still operate. This
means that, unlike Apache, there is a "default" server configuration
proftpd.conf; ProFTPD does not require that all
server configurations be explicitly written in the
file. This default server attempts to bind to the IP address of the hostname
indicated by the
The format of the
proftpd.conf file is deliberately designed
to resemble the format used by Apache: lines of configuration directives
contained with different contexts. A list of the configuration directives
for ProFTPD is available here:
http://www.proftpd.org/docs/When reading the description for the configuration directives, this key might be useful:
http://www.castaglia.org/proftpd/doc/contrib/configuration-directive-key.htmlIt describes the description format, and lists the different contexts in the configuration file. The "server config" context is the one in which most of your configuration directives will most likely be placed.
Two new configuration directives were introduced in
work exactly like Apache's directives of the same names, providing the ability
to have conditional sections in the configuration file.
Starting the Daemon
One of the first decisions you will need to make is whether you will be running your ProFTPD server as an
inetd service, or as a
standalone server. This is controlled in the
proftpd.conf using the
directive (see the ServerType page).
The daemon must be started with root privileges in order to do things like binding to port 21 and chrooting FTP sessions. However, it is not a good idea to leave a long-lived process running as root. The
Group configuration directives are thus recommended. These
directives configure the identity to which the daemon will switch, after
accomplishing its startup tasks. The daemon will switch to the configured
Group in the "server config"
context. (Note that this switch uses the effective UID/GID,
not the real UID/GID. Some programs, such as
will continue to report
proftpd as running as
this is because the program displays the real UID/GID of processes.
proftpd daemon retains root privileges for operations
such as chroots and binding to port 20 for active data transfers.
If you wish
proftpd to drop all root privileges, use the
RootRevoke configuration directive.)
For this reason, it is recommended that a non-privileged identity be
used. Many sites choose to use user
this role account was used by NFS-related processes; over time, many other
applications default to using user
nobody. This has the
side-effect of adding to the "privileges" held by user
nobody, in terms of files owned and/or accessible by that user.
Instead, I personally recommend that a new role account be created for use
specifically by the daemon, a user
ftpd, and perhaps even a
ftpd. Many systems that run Apache have a user
www or user
apache for use by the
daemon; similarly, a separate user should be created for the
In the default configuration file that accompanies the proftpd source code, there appears:
User nobody Group nogroupWhen trying to start the daemon, many users encounter the "no such group 'nogroup'" error message. There are really no reasonable defaults for those directives. The error message is a way of telling you to create the role accounts mentioned above.
By default, the
proftpd daemon reads the host's
/etc/passwd file for logging in users. This means that to
add FTP users, you simply need to create new system accounts for those users in
Sometimes, though, sites want "virtual", FTP-only users. In order
to support such configurations, the
directive can be used (see here for details).
For the purpose of authenticating users using other means, there are various
mod_radius, etc. Authentication and the login process
is discussed here in more detail.
For setting up anonymous logins, there is the
<Anonymous> configuration context. If there are no
sections in your
proftpd.conf, then no anonymous logins will be
allowed - simple. As mentioned in the description, the
directive in an
<Anonymous> context determines what username
is treated as an anonymous login. The main other thing to know about
anonymous logins is that ProFTPD automatically chroots anonymous logins.
For normal, non-anonymous logins, jails/chroots are configured using the
DefaultRoot directive. This is the configuration directive
used to restrict users to their home directories, to keep them from browsing
around the site. There is a page covering chrooting
If you use
<VirtualHost> sections, and it seems that your
server configuration is not being seen by connecting clients, you might
need to check that, if using a DNS name instead of an IP address in your
<VirtualHost> line, that name resolves to an IP address
different from that of the "default" server. Many people new
to ProFTPD get the impression that since the configuration syntax looks
similar to Apache's, things like name-based virtual hosting will work as well.
Unfortunately, this is not possible. It is not a limitation in ProFTPD,
but rather in the RFCs that define FTP. See the
virtual server page for more information.
As a workaround, some sites configure virtual servers to run on non-standard
ports, using the
Port configuration directive. As long as
the clients are aware of the non-standard port, this scheme works well. One
minor little caveat to keep in mind, when using this approach, is the numbers
used: the RFCs mandate that the daemon, for the purposes of active data
transfers (as opposed to passive) use port
L-1 as the source
port for the data connection, where
L is the port number
at which the client contacted the server. This means that servers that use the
standard port 21 for FTP will use port 20 as the source port for their active
data transfers. Passive data transfers do not have this restriction. The
restriction comes into play when choosing non-standard port numbers for
virtual hosts. For example, this configuration would cause problems for
clients of the second virtual server that wanted to use active data transfers:
<VirtualHost a.b.c.d> Port 2121 ... </VirtualHost> <VirtualHost a.b.c.d> Port 2122 ... </VirtualHost>The second virtual would attempt to use port 2121 as the source port for an active data transfer, but would be blocked, as the first virtual server is already using that port for listening.
Many sites like to have specific directories for uploads, and other directories only for downloads; some sites like to allow downloads, but no browsing of directories or their contents. For configurations to achieve this, use combinations of the
<Limit> configuration directives. There are separate
pages that cover these configuration sections:
Hopefully this document answers some of your questions, or at least enough to get you started. In addition, you should take a look at some of the example configuration files. Once you are comfortable with the configuration file format, a reading of all the configuration directives' descriptions is recommended, especially if you plan on having more complex configurations. When trying to figure out why something is not working, make use of server debugging output.
If you still have questions, the users mailing list is the best place to post them.