2008-11-20 12:48:26 +01:00
<!-- $PostgreSQL: pgsql/doc/src/sgml/client-auth.sgml,v 1.112 2008/11/20 11:48:26 mha Exp $ -->
2000-06-18 23:24:54 +02:00
<chapter id="client-authentication">
<title>Client Authentication</title>
2001-05-13 00:51:36 +02:00
<indexterm zone="client-authentication">
<primary>client authentication</primary>
</indexterm>
2000-06-18 23:24:54 +02:00
<para>
2002-04-04 06:25:54 +02:00
When a client application connects to the database server, it
2005-08-15 01:35:38 +02:00
specifies which <productname>PostgreSQL</productname> database user name it
2002-04-04 06:25:54 +02:00
wants to connect as, much the same way one logs into a Unix computer
as a particular user. Within the SQL environment the active database
2004-11-15 07:32:15 +01:00
user name determines access privileges to database objects — see
2002-04-04 06:25:54 +02:00
<xref linkend="user-manag"> for more information. Therefore, it is
essential to restrict which database users can connect.
2000-06-18 23:24:54 +02:00
</para>
2005-08-15 01:35:38 +02:00
<note>
<para>
As explained in <xref linkend="user-manag">,
<productname>PostgreSQL</productname> actually does privilege
management in terms of <quote>roles</>. In this chapter, we
consistently use <firstterm>database user</> to mean <quote>role with the
<literal>LOGIN</> privilege</quote>.
</para>
</note>
2000-06-18 23:24:54 +02:00
<para>
<firstterm>Authentication</firstterm> is the process by which the
database server establishes the identity of the client, and by
extension determines whether the client application (or the user
2000-11-21 21:44:32 +01:00
who runs the client application) is permitted to connect with the
2005-08-15 01:35:38 +02:00
database user name that was requested.
2000-06-18 23:24:54 +02:00
</para>
<para>
2001-11-21 06:53:41 +01:00
<productname>PostgreSQL</productname> offers a number of different
2002-08-16 06:48:16 +02:00
client authentication methods. The method used to authenticate a
particular client connection can be selected on the basis of
(client) host address, database, and user.
2000-06-18 23:24:54 +02:00
</para>
2000-11-21 21:44:32 +01:00
<para>
2005-08-15 01:35:38 +02:00
<productname>PostgreSQL</productname> database user names are logically
2000-11-21 21:44:32 +01:00
separate from user names of the operating system in which the server
2002-04-04 06:25:54 +02:00
runs. If all the users of a particular server also have accounts on
2000-11-21 21:44:32 +01:00
the server's machine, it makes sense to assign database user names
2002-04-04 06:25:54 +02:00
that match their operating system user names. However, a server that
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
accepts remote connections might have many database users who have no local operating system
2002-04-04 06:25:54 +02:00
account, and in such cases there need be no connection between
database user names and OS user names.
2000-11-21 21:44:32 +01:00
</para>
2002-09-14 20:35:46 +02:00
<sect1 id="auth-pg-hba-conf">
2000-06-18 23:24:54 +02:00
<title>The <filename>pg_hba.conf</filename> file</title>
2002-09-14 20:35:46 +02:00
<indexterm zone="auth-pg-hba-conf">
2001-05-13 00:51:36 +02:00
<primary>pg_hba.conf</primary>
</indexterm>
2000-06-18 23:24:54 +02:00
<para>
2004-12-27 00:06:56 +01:00
Client authentication is controlled by a configuration file,
which traditionally is named
<filename>pg_hba.conf</filename> and is stored in the database
cluster's data directory.
2002-04-04 06:25:54 +02:00
(<acronym>HBA</> stands for host-based authentication.) A default
2002-08-16 06:48:16 +02:00
<filename>pg_hba.conf</filename> file is installed when the data
2004-12-27 00:06:56 +01:00
directory is initialized by <command>initdb</command>. It is
possible to place the authentication configuration file elsewhere,
however; see the <xref linkend="guc-hba-file"> configuration parameter.
2000-06-18 23:24:54 +02:00
</para>
<para>
2002-11-11 21:14:04 +01:00
The general format of the <filename>pg_hba.conf</filename> file is
2002-04-04 06:25:54 +02:00
a set of records, one per line. Blank lines are ignored, as is any
2003-03-13 02:30:29 +01:00
text after the <literal>#</literal> comment character. A record is made
2002-04-04 06:25:54 +02:00
up of a number of fields which are separated by spaces and/or tabs.
Fields can contain white space if the field value is quoted. Records
cannot be continued across lines.
2000-06-18 23:24:54 +02:00
</para>
2001-11-02 19:39:57 +01:00
<para>
Each record specifies a connection type, a client IP address range
2002-04-04 06:25:54 +02:00
(if relevant for the connection type), a database name, a user name,
2001-11-02 19:39:57 +01:00
and the authentication method to be used for connections matching
2002-04-04 06:25:54 +02:00
these parameters. The first record with a matching connection type,
client address, requested database, and user name is used to perform
authentication. There is no <quote>fall-through</> or
2001-11-02 19:39:57 +01:00
<quote>backup</>: if one record is chosen and the authentication
2002-04-04 06:25:54 +02:00
fails, subsequent records are not considered. If no record matches,
access is denied.
2001-11-02 19:39:57 +01:00
</para>
2000-06-18 23:24:54 +02:00
<para>
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
A record can have one of the seven formats
2003-03-13 02:30:29 +01:00
<synopsis>
2008-10-23 15:31:10 +02:00
local <replaceable>database</replaceable> <replaceable>user</replaceable> <replaceable>auth-method</replaceable> <optional><replaceable>auth-options</replaceable></optional>
host <replaceable>database</replaceable> <replaceable>user</replaceable> <replaceable>CIDR-address</replaceable> <replaceable>auth-method</replaceable> <optional><replaceable>auth-options</replaceable></optional>
hostssl <replaceable>database</replaceable> <replaceable>user</replaceable> <replaceable>CIDR-address</replaceable> <replaceable>auth-method</replaceable> <optional><replaceable>auth-options</replaceable></optional>
hostnossl <replaceable>database</replaceable> <replaceable>user</replaceable> <replaceable>CIDR-address</replaceable> <replaceable>auth-method</replaceable> <optional><replaceable>auth-options</replaceable></optional>
host <replaceable>database</replaceable> <replaceable>user</replaceable> <replaceable>IP-address</replaceable> <replaceable>IP-mask</replaceable> <replaceable>auth-method</replaceable> <optional><replaceable>auth-options</replaceable></optional>
hostssl <replaceable>database</replaceable> <replaceable>user</replaceable> <replaceable>IP-address</replaceable> <replaceable>IP-mask</replaceable> <replaceable>auth-method</replaceable> <optional><replaceable>auth-options</replaceable></optional>
hostnossl <replaceable>database</replaceable> <replaceable>user</replaceable> <replaceable>IP-address</replaceable> <replaceable>IP-mask</replaceable> <replaceable>auth-method</replaceable> <optional><replaceable>auth-options</replaceable></optional>
2003-03-13 02:30:29 +01:00
</synopsis>
2000-06-18 23:24:54 +02:00
The meaning of the fields is as follows:
<variablelist>
<varlistentry>
<term><literal>local</literal></term>
<listitem>
<para>
2003-03-13 02:30:29 +01:00
This record matches connection attempts using Unix-domain
2002-09-14 20:35:46 +02:00
sockets. Without a record of this type, Unix-domain socket
2003-09-13 18:43:38 +02:00
connections are disallowed.
2000-06-18 23:24:54 +02:00
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>host</literal></term>
<listitem>
<para>
2003-12-14 00:59:07 +01:00
This record matches connection attempts made using TCP/IP.
2004-03-23 02:23:48 +01:00
<literal>host</literal> records match either
2003-12-14 00:59:07 +01:00
<acronym>SSL</acronym> or non-<acronym>SSL</acronym> connection
attempts.
2000-06-18 23:24:54 +02:00
</para>
2004-03-23 02:23:48 +01:00
<note>
<para>
Remote TCP/IP connections will not be possible unless
the server is started with an appropriate value for the
<xref linkend="guc-listen-addresses"> configuration parameter,
since the default behavior is to listen for TCP/IP connections
only on the local loopback address <literal>localhost</>.
</para>
</note>
2000-06-18 23:24:54 +02:00
</listitem>
</varlistentry>
2000-08-29 06:15:43 +02:00
<varlistentry>
<term><literal>hostssl</literal></term>
<listitem>
<para>
2004-03-23 02:23:48 +01:00
This record matches connection attempts made using TCP/IP,
but only when the connection is made with <acronym>SSL</acronym>
encryption.
2002-09-14 20:35:46 +02:00
</para>
<para>
2003-12-14 00:59:07 +01:00
To make use of this option the server must be built with
2004-03-23 02:23:48 +01:00
<acronym>SSL</acronym> support. Furthermore,
<acronym>SSL</acronym> must be enabled at server start time
by setting the <xref linkend="guc-ssl"> configuration parameter (see
<xref linkend="ssl-tcp"> for more information).
2000-08-29 06:15:43 +02:00
</para>
</listitem>
</varlistentry>
At long last I put together a patch to support 4 client SSL negotiation
modes (and replace the requiressl boolean). The four options were first
spelled out by Magnus Hagander <mha@sollentuna.net> on 2000-08-23 in email
to pgsql-hackers, archived here:
http://archives.postgresql.org/pgsql-hackers/2000-08/msg00639.php
My original less-flexible patch and the ensuing thread are archived at:
http://dbforums.com/t623845.html
Attached is a new patch, including documentation.
To sum up, there's a new client parameter "sslmode" and environment
variable "PGSSLMODE", with these options:
sslmode description
------- -----------
disable Unencrypted non-SSL only
allow Negotiate, prefer non-SSL
prefer Negotiate, prefer SSL (default)
require Require SSL
The only change to the server is a new pg_hba.conf line type,
"hostnossl", for specifying connections that are not allowed to use SSL
(for example, to prevent servers on a local network from accidentally
using SSL and wasting cycles). Thus the 3 pg_hba.conf line types are:
pg_hba.conf line types
----------------------
host applies to either SSL or regular connections
hostssl applies only to SSL connections
hostnossl applies only to regular connections
These client and server options, the postgresql.conf ssl = false option,
and finally the possibility of compiling with no SSL support at all,
make quite a range of combinations to test. I threw together a test
script to try many of them out. It's in a separate tarball with its
config files, a patch to psql so it'll announce SSL connections even in
absence of a tty, and the test output. The test is especially informative
when run on the same tty the postmaster was started on, so the FATAL:
errors during negotiation are interleaved with the psql client output.
I saw Tom write that new submissions for 7.4 have to be in before midnight
local time, and since I'm on the east coast in the US, this just makes it
in before the bell. :)
Jon Jensen
2003-07-26 15:50:02 +02:00
<varlistentry>
<term><literal>hostnossl</literal></term>
<listitem>
<para>
2004-12-27 00:06:56 +01:00
This record type has the opposite logic to <literal>hostssl</>:
it only matches connection attempts made over
2004-03-23 02:23:48 +01:00
TCP/IP that do not use <acronym>SSL</acronym>.
At long last I put together a patch to support 4 client SSL negotiation
modes (and replace the requiressl boolean). The four options were first
spelled out by Magnus Hagander <mha@sollentuna.net> on 2000-08-23 in email
to pgsql-hackers, archived here:
http://archives.postgresql.org/pgsql-hackers/2000-08/msg00639.php
My original less-flexible patch and the ensuing thread are archived at:
http://dbforums.com/t623845.html
Attached is a new patch, including documentation.
To sum up, there's a new client parameter "sslmode" and environment
variable "PGSSLMODE", with these options:
sslmode description
------- -----------
disable Unencrypted non-SSL only
allow Negotiate, prefer non-SSL
prefer Negotiate, prefer SSL (default)
require Require SSL
The only change to the server is a new pg_hba.conf line type,
"hostnossl", for specifying connections that are not allowed to use SSL
(for example, to prevent servers on a local network from accidentally
using SSL and wasting cycles). Thus the 3 pg_hba.conf line types are:
pg_hba.conf line types
----------------------
host applies to either SSL or regular connections
hostssl applies only to SSL connections
hostnossl applies only to regular connections
These client and server options, the postgresql.conf ssl = false option,
and finally the possibility of compiling with no SSL support at all,
make quite a range of combinations to test. I threw together a test
script to try many of them out. It's in a separate tarball with its
config files, a patch to psql so it'll announce SSL connections even in
absence of a tty, and the test output. The test is especially informative
when run on the same tty the postmaster was started on, so the FATAL:
errors during negotiation are interleaved with the psql client output.
I saw Tom write that new submissions for 7.4 have to be in before midnight
local time, and since I'm on the east coast in the US, this just makes it
in before the bell. :)
Jon Jensen
2003-07-26 15:50:02 +02:00
</para>
</listitem>
</varlistentry>
2000-06-18 23:24:54 +02:00
<varlistentry>
<term><replaceable>database</replaceable></term>
<listitem>
<para>
2005-08-15 01:35:38 +02:00
Specifies which database names this record matches. The value
2002-09-14 20:35:46 +02:00
<literal>all</literal> specifies that it matches all databases.
The value <literal>sameuser</> specifies that the record
matches if the requested database has the same name as the
2005-08-15 01:35:38 +02:00
requested user. The value <literal>samerole</> specifies that
the requested user must be a member of the role with the same
name as the requested database. (<literal>samegroup</> is an
obsolete but still accepted spelling of <literal>samerole</>.)
Otherwise, this is the name of
2002-09-14 20:35:46 +02:00
a specific <productname>PostgreSQL</productname> database.
Multiple database names can be supplied by separating them with
2005-08-15 01:35:38 +02:00
commas. A separate file containing database names can be specified by
2004-12-27 20:19:24 +01:00
preceding the file name with <literal>@</>.
2002-04-04 06:25:54 +02:00
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><replaceable>user</replaceable></term>
<listitem>
<para>
2005-08-15 01:35:38 +02:00
Specifies which database user names this record
2002-11-15 04:11:18 +01:00
matches. The value <literal>all</literal> specifies that it
2005-08-15 01:35:38 +02:00
matches all users. Otherwise, this is either the name of a specific
database user, or a group name preceded by <literal>+</>.
(Recall that there is no real distinction between users and groups
in <productname>PostgreSQL</>; a <literal>+</> mark really means
<quote>match any of the roles that are directly or indirectly members
of this role</>, while a name without a <literal>+</> mark matches
only that specific role.)
Multiple user names can be supplied by separating them with commas.
A separate file containing user names can be specified by preceding the
2004-12-27 20:19:24 +01:00
file name with <literal>@</>.
2000-06-18 23:24:54 +02:00
</para>
</listitem>
</varlistentry>
<varlistentry>
2004-08-26 18:50:05 +02:00
<term><replaceable>CIDR-address</replaceable></term>
2000-06-18 23:24:54 +02:00
<listitem>
<para>
2004-12-27 00:06:56 +01:00
Specifies the client machine IP address range that this record
2004-08-26 18:50:05 +02:00
matches. It contains an IP address in standard dotted decimal
notation and a CIDR mask length. (IP addresses can only be
2004-12-27 00:06:56 +01:00
specified numerically, not as domain or host names.) The mask
length indicates the number of high-order bits of the client
IP address that must match. Bits to the right of this must
be zero in the given IP address.
There must not be any white space between the IP address, the
<literal>/</literal>, and the CIDR mask length.
2004-08-26 18:50:05 +02:00
</para>
<para>
2006-11-18 00:25:17 +01:00
Typical examples of a <replaceable>CIDR-address</replaceable> are
2004-12-27 00:06:56 +01:00
<literal>172.20.143.89/32</literal> for a single host, or
2006-11-18 00:25:17 +01:00
<literal>172.20.143.0/24</literal> for a small network, or
<literal>10.6.0.0/16</literal> for a larger one.
2004-12-27 00:06:56 +01:00
To specify a single host, use a CIDR mask of 32 for IPv4 or
2006-11-18 00:25:17 +01:00
128 for IPv6. In a network address, do not omit trailing zeroes.
2003-09-05 22:31:36 +02:00
</para>
<para>
An IP address given in IPv4 format will match IPv6 connections that
have the corresponding address, for example <literal>127.0.0.1</>
will match the IPv6 address <literal>::ffff:127.0.0.1</>. An entry
given in IPv6 format will match only IPv6 connections, even if the
represented address is in the IPv4-in-IPv6 range. Note that entries
in IPv6 format will be rejected if the system's C library does not have
support for IPv6 addresses.
2002-09-14 20:35:46 +02:00
</para>
<para>
2004-12-27 00:06:56 +01:00
This field only applies to <literal>host</literal>,
At long last I put together a patch to support 4 client SSL negotiation
modes (and replace the requiressl boolean). The four options were first
spelled out by Magnus Hagander <mha@sollentuna.net> on 2000-08-23 in email
to pgsql-hackers, archived here:
http://archives.postgresql.org/pgsql-hackers/2000-08/msg00639.php
My original less-flexible patch and the ensuing thread are archived at:
http://dbforums.com/t623845.html
Attached is a new patch, including documentation.
To sum up, there's a new client parameter "sslmode" and environment
variable "PGSSLMODE", with these options:
sslmode description
------- -----------
disable Unencrypted non-SSL only
allow Negotiate, prefer non-SSL
prefer Negotiate, prefer SSL (default)
require Require SSL
The only change to the server is a new pg_hba.conf line type,
"hostnossl", for specifying connections that are not allowed to use SSL
(for example, to prevent servers on a local network from accidentally
using SSL and wasting cycles). Thus the 3 pg_hba.conf line types are:
pg_hba.conf line types
----------------------
host applies to either SSL or regular connections
hostssl applies only to SSL connections
hostnossl applies only to regular connections
These client and server options, the postgresql.conf ssl = false option,
and finally the possibility of compiling with no SSL support at all,
make quite a range of combinations to test. I threw together a test
script to try many of them out. It's in a separate tarball with its
config files, a patch to psql so it'll announce SSL connections even in
absence of a tty, and the test output. The test is especially informative
when run on the same tty the postmaster was started on, so the FATAL:
errors during negotiation are interleaved with the psql client output.
I saw Tom write that new submissions for 7.4 have to be in before midnight
local time, and since I'm on the east coast in the US, this just makes it
in before the bell. :)
Jon Jensen
2003-07-26 15:50:02 +02:00
<literal>hostssl</literal>, and <literal>hostnossl</> records.
2000-06-18 23:24:54 +02:00
</para>
</listitem>
</varlistentry>
2003-06-25 03:20:50 +02:00
<varlistentry>
2004-08-26 18:50:05 +02:00
<term><replaceable>IP-address</replaceable></term>
2004-12-27 00:06:56 +01:00
<term><replaceable>IP-mask</replaceable></term>
2003-06-25 03:20:50 +02:00
<listitem>
<para>
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
These fields can be used as an alternative to the
2004-08-26 18:50:05 +02:00
<replaceable>CIDR-address</replaceable> notation. Instead of
specifying the mask length, the actual mask is specified in a
2004-12-27 00:06:56 +01:00
separate column. For example, <literal>255.0.0.0</> represents an IPv4
CIDR mask length of 8, and <literal>255.255.255.255</> represents a
CIDR mask length of 32.
2003-06-25 03:20:50 +02:00
</para>
<para>
2004-12-27 00:06:56 +01:00
These fields only apply to <literal>host</literal>,
At long last I put together a patch to support 4 client SSL negotiation
modes (and replace the requiressl boolean). The four options were first
spelled out by Magnus Hagander <mha@sollentuna.net> on 2000-08-23 in email
to pgsql-hackers, archived here:
http://archives.postgresql.org/pgsql-hackers/2000-08/msg00639.php
My original less-flexible patch and the ensuing thread are archived at:
http://dbforums.com/t623845.html
Attached is a new patch, including documentation.
To sum up, there's a new client parameter "sslmode" and environment
variable "PGSSLMODE", with these options:
sslmode description
------- -----------
disable Unencrypted non-SSL only
allow Negotiate, prefer non-SSL
prefer Negotiate, prefer SSL (default)
require Require SSL
The only change to the server is a new pg_hba.conf line type,
"hostnossl", for specifying connections that are not allowed to use SSL
(for example, to prevent servers on a local network from accidentally
using SSL and wasting cycles). Thus the 3 pg_hba.conf line types are:
pg_hba.conf line types
----------------------
host applies to either SSL or regular connections
hostssl applies only to SSL connections
hostnossl applies only to regular connections
These client and server options, the postgresql.conf ssl = false option,
and finally the possibility of compiling with no SSL support at all,
make quite a range of combinations to test. I threw together a test
script to try many of them out. It's in a separate tarball with its
config files, a patch to psql so it'll announce SSL connections even in
absence of a tty, and the test output. The test is especially informative
when run on the same tty the postmaster was started on, so the FATAL:
errors during negotiation are interleaved with the psql client output.
I saw Tom write that new submissions for 7.4 have to be in before midnight
local time, and since I'm on the east coast in the US, this just makes it
in before the bell. :)
Jon Jensen
2003-07-26 15:50:02 +02:00
<literal>hostssl</literal>, and <literal>hostnossl</> records.
2003-06-25 03:20:50 +02:00
</para>
</listitem>
</varlistentry>
2000-06-18 23:24:54 +02:00
<varlistentry>
2005-08-15 01:35:38 +02:00
<term><replaceable>auth-method</replaceable></term>
2000-06-18 23:24:54 +02:00
<listitem>
<para>
2002-04-04 06:25:54 +02:00
Specifies the authentication method to use when connecting via
this record. The possible choices are summarized here; details
are in <xref linkend="auth-methods">.
2000-09-06 21:54:52 +02:00
<variablelist>
<varlistentry>
2001-09-10 01:52:12 +02:00
<term><literal>trust</></term>
2000-09-06 21:54:52 +02:00
<listitem>
<para>
2004-12-27 00:06:56 +01:00
Allow the connection unconditionally. This method
2002-08-16 06:48:16 +02:00
allows anyone that can connect to the
2003-03-13 02:30:29 +01:00
<productname>PostgreSQL</productname> database server to login as
2002-08-16 06:48:16 +02:00
any <productname>PostgreSQL</productname> user they like,
2002-09-14 20:35:46 +02:00
without the need for a password. See <xref
linkend="auth-trust"> for details.
2000-09-06 21:54:52 +02:00
</para>
</listitem>
</varlistentry>
<varlistentry>
2001-09-10 01:52:12 +02:00
<term><literal>reject</></term>
2000-09-06 21:54:52 +02:00
<listitem>
<para>
2004-12-27 00:06:56 +01:00
Reject the connection unconditionally. This is useful for
2002-04-04 06:25:54 +02:00
<quote>filtering out</> certain hosts from a group.
2000-09-06 21:54:52 +02:00
</para>
</listitem>
</varlistentry>
<varlistentry>
2002-04-04 06:25:54 +02:00
<term><literal>md5</></term>
2000-09-06 21:54:52 +02:00
<listitem>
<para>
2004-12-27 00:06:56 +01:00
Require the client to supply an MD5-encrypted password for
authentication.
2002-09-14 20:35:46 +02:00
See <xref linkend="auth-password"> for details.
2000-09-06 21:54:52 +02:00
</para>
</listitem>
</varlistentry>
2001-08-16 18:24:16 +02:00
<varlistentry>
2002-04-04 06:25:54 +02:00
<term><literal>password</></term>
2001-08-16 18:24:16 +02:00
<listitem>
<para>
2004-12-27 00:06:56 +01:00
Require the client to supply an unencrypted password for
authentication.
Since the password is sent in clear text over the
network, this should not be used on untrusted networks.
2005-10-24 17:38:37 +02:00
It also does not usually work with threaded client applications.
2002-09-14 20:35:46 +02:00
See <xref linkend="auth-password"> for details.
2002-04-04 06:25:54 +02:00
</para>
2001-08-16 18:24:16 +02:00
</listitem>
</varlistentry>
2007-07-18 14:00:47 +02:00
<varlistentry>
<term><literal>gss</></term>
<listitem>
<para>
Use GSSAPI to authenticate the user. This is only
available for TCP/IP connections. See <xref
linkend="gssapi-auth"> for details.
</para>
</listitem>
</varlistentry>
2007-07-23 12:16:54 +02:00
<varlistentry>
<term><literal>sspi</></term>
<listitem>
<para>
Use SSPI to authenticate the user. This is only
available on Windows. See <xref
linkend="sspi-auth"> for details.
</para>
</listitem>
</varlistentry>
2000-09-06 21:54:52 +02:00
<varlistentry>
2001-09-10 01:52:12 +02:00
<term><literal>krb5</></term>
2000-09-06 21:54:52 +02:00
<listitem>
<para>
2004-12-27 00:06:56 +01:00
Use Kerberos V5 to authenticate the user. This is only
2002-09-14 20:35:46 +02:00
available for TCP/IP connections. See <xref
linkend="kerberos-auth"> for details.
2000-09-06 21:54:52 +02:00
</para>
</listitem>
</varlistentry>
<varlistentry>
2001-09-10 01:52:12 +02:00
<term><literal>ident</></term>
2000-09-06 21:54:52 +02:00
<listitem>
2002-09-14 20:35:46 +02:00
<para>
Obtain the operating system user name of the client (for
TCP/IP connections by contacting the ident server on the
client, for local connections by getting it from the
operating system) and check if the user is allowed to
connect as the requested database user by consulting the map
specified after the <literal>ident</literal> key word.
2004-12-27 00:06:56 +01:00
See <xref linkend="auth-ident"> for details.
2000-09-06 21:54:52 +02:00
</para>
</listitem>
</varlistentry>
2001-09-06 05:23:38 +02:00
2006-06-16 17:16:16 +02:00
<varlistentry>
<term><literal>ldap</></term>
<listitem>
<para>
Authenticate using LDAP to a central server. See <xref
linkend="auth-ldap"> for details.
</para>
</listitem>
</varlistentry>
2008-11-20 12:48:26 +01:00
<varlistentry>
<term><literal>cert</></term>
<listitem>
<para>
Authenticate using SSL client certificates. See
<xref linkend="auth-cert"> for details.
</para>
</listitem>
</varlistentry>
2001-09-06 05:23:38 +02:00
<varlistentry>
2001-09-10 01:52:12 +02:00
<term><literal>pam</></term>
2001-09-06 05:23:38 +02:00
<listitem>
<para>
2002-09-14 20:35:46 +02:00
Authenticate using the Pluggable Authentication Modules
(PAM) service provided by the operating system. See <xref
linkend="auth-pam"> for details.
2001-09-06 05:23:38 +02:00
</para>
</listitem>
</varlistentry>
2000-09-06 21:54:52 +02:00
</variablelist>
2000-06-18 23:24:54 +02:00
</para>
</listitem>
</varlistentry>
2002-09-14 20:35:46 +02:00
<varlistentry>
2008-10-23 15:31:10 +02:00
<term><replaceable>auth-options</replaceable></term>
2002-09-14 20:35:46 +02:00
<listitem>
<para>
2008-10-23 15:31:10 +02:00
This field contains zero or more name-value pairs with
extra options passed to this authentication method. Details
about which options are available for which authentication
method appear below.
2002-09-14 20:35:46 +02:00
</para>
</listitem>
</varlistentry>
2000-06-18 23:24:54 +02:00
</variablelist>
2001-11-02 19:39:57 +01:00
</para>
2000-06-18 23:24:54 +02:00
2001-11-02 19:39:57 +01:00
<para>
2004-12-27 20:19:24 +01:00
Files included by <literal>@</> constructs are read as lists of names,
which can be separated by either whitespace or commas. Comments are
introduced by <literal>#</literal>, just as in
<filename>pg_hba.conf</filename>, and nested <literal>@</> constructs are
allowed. Unless the file name following <literal>@</> is an absolute
path, it is taken to be relative to the directory containing the
referencing file.
</para>
<para>
2001-11-02 19:39:57 +01:00
Since the <filename>pg_hba.conf</filename> records are examined
2001-11-28 21:49:10 +01:00
sequentially for each connection attempt, the order of the records is
2002-04-04 06:25:54 +02:00
significant. Typically, earlier records will have tight connection
match parameters and weaker authentication methods, while later
records will have looser match parameters and stronger authentication
methods. For example, one might wish to use <literal>trust</>
2003-03-13 02:30:29 +01:00
authentication for local TCP/IP connections but require a password for
remote TCP/IP connections. In this case a record specifying
2002-04-04 06:25:54 +02:00
<literal>trust</> authentication for connections from 127.0.0.1 would
appear before a record specifying password authentication for a wider
range of allowed client IP addresses.
2000-06-18 23:24:54 +02:00
</para>
<para>
2002-04-04 06:25:54 +02:00
The <filename>pg_hba.conf</filename> file is read on start-up and when
2006-06-18 17:38:37 +02:00
the main server process receives a
2003-03-13 02:30:29 +01:00
<systemitem>SIGHUP</systemitem><indexterm><primary>SIGHUP</primary></indexterm>
signal. If you edit the file on an
2006-06-18 17:38:37 +02:00
active system, you will need to signal the server
2002-04-04 06:25:54 +02:00
(using <literal>pg_ctl reload</> or <literal>kill -HUP</>) to make it
re-read the file.
2000-06-18 23:24:54 +02:00
</para>
2006-04-30 23:15:33 +02:00
<tip>
<para>
To connect to a particular database, a user must not only pass the
<filename>pg_hba.conf</filename> checks, but must have the
<literal>CONNECT</> privilege for the database. If you wish to
restrict which users can connect to which databases, it's usually
easier to control this by granting/revoking <literal>CONNECT</> privilege
than to put the rules into <filename>pg_hba.conf</filename> entries.
</para>
</tip>
2000-06-18 23:24:54 +02:00
<para>
2004-12-27 00:06:56 +01:00
Some examples of <filename>pg_hba.conf</filename> entries are shown in
2003-03-13 02:30:29 +01:00
<xref linkend="example-pg-hba.conf">. See the next section for details on the
2000-06-18 23:24:54 +02:00
different authentication methods.
2003-03-13 02:30:29 +01:00
</para>
2000-06-18 23:24:54 +02:00
<example id="example-pg-hba.conf">
2004-12-27 00:06:56 +01:00
<title>Example <filename>pg_hba.conf</filename> entries</title>
2000-06-18 23:24:54 +02:00
<programlisting>
2002-09-14 20:35:46 +02:00
# Allow any user on the local system to connect to any database under
2005-08-15 01:35:38 +02:00
# any database user name using Unix-domain sockets (the default for local
2002-09-14 20:35:46 +02:00
# connections).
#
2004-08-26 18:50:05 +02:00
# TYPE DATABASE USER CIDR-ADDRESS METHOD
local all all trust
2002-09-14 20:35:46 +02:00
# The same using local loopback TCP/IP connections.
#
2004-08-26 18:50:05 +02:00
# TYPE DATABASE USER CIDR-ADDRESS METHOD
host all all 127.0.0.1/32 trust
2002-09-14 20:35:46 +02:00
2004-08-26 18:50:05 +02:00
# The same as the last line but using a separate netmask column
2003-08-17 06:39:11 +02:00
#
2004-12-27 00:06:56 +01:00
# TYPE DATABASE USER IP-ADDRESS IP-MASK METHOD
2004-08-26 18:50:05 +02:00
host all all 127.0.0.1 255.255.255.255 trust
2003-08-17 06:39:11 +02:00
2002-09-14 20:35:46 +02:00
# Allow any user from any host with IP address 192.168.93.x to connect
2005-06-21 06:02:34 +02:00
# to database "postgres" as the same user name that ident reports for
2002-09-14 20:35:46 +02:00
# the connection (typically the Unix user name).
#
2004-08-26 18:50:05 +02:00
# TYPE DATABASE USER CIDR-ADDRESS METHOD
2008-09-15 14:32:57 +02:00
host postgres all 192.168.93.0/24 ident
2002-09-14 20:35:46 +02:00
# Allow a user from host 192.168.12.10 to connect to database
2005-06-21 06:02:34 +02:00
# "postgres" if the user's password is correctly supplied.
2002-09-14 20:35:46 +02:00
#
2004-08-26 18:50:05 +02:00
# TYPE DATABASE USER CIDR-ADDRESS METHOD
2005-06-21 06:02:34 +02:00
host postgres all 192.168.12.10/32 md5
2002-09-14 20:35:46 +02:00
# In the absence of preceding "host" lines, these two lines will
# reject all connection from 192.168.54.1 (since that entry will be
2004-12-27 00:06:56 +01:00
# matched first), but allow Kerberos 5 connections from anywhere else
2002-09-14 20:35:46 +02:00
# on the Internet. The zero mask means that no bits of the host IP
# address are considered so it matches any host.
#
2004-08-26 18:50:05 +02:00
# TYPE DATABASE USER CIDR-ADDRESS METHOD
host all all 192.168.54.1/32 reject
host all all 0.0.0.0/0 krb5
2002-09-14 20:35:46 +02:00
# Allow users from 192.168.x.x hosts to connect to any database, if
# they pass the ident check. If, for example, ident says the user is
# "bryanh" and he requests to connect as PostgreSQL user "guest1", the
# connection is allowed if there is an entry in pg_ident.conf for map
# "omicron" that says "bryanh" is allowed to connect as "guest1".
#
2004-08-26 18:50:05 +02:00
# TYPE DATABASE USER CIDR-ADDRESS METHOD
2008-10-23 15:31:10 +02:00
host all all 192.168.0.0/16 ident map=omicron
2002-09-14 20:35:46 +02:00
# If these are the only three lines for local connections, they will
# allow local users to connect only to their own databases (databases
2005-08-15 01:35:38 +02:00
# with the same name as their database user name) except for administrators
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
# and members of role "support", who can connect to all databases. The file
2005-08-15 01:35:38 +02:00
# $PGDATA/admins contains a list of names of administrators. Passwords
# are required in all cases.
2002-09-14 20:35:46 +02:00
#
2004-08-26 18:50:05 +02:00
# TYPE DATABASE USER CIDR-ADDRESS METHOD
2004-11-10 22:54:23 +01:00
local sameuser all md5
local all @admins md5
local all +support md5
2002-09-14 20:35:46 +02:00
# The last two lines above can be combined into a single line:
2004-11-10 22:54:23 +01:00
local all @admins,+support md5
2002-09-14 20:35:46 +02:00
2005-08-15 01:35:38 +02:00
# The database column can also use lists and file names:
2004-11-10 22:54:23 +01:00
local db1,db2,@demodbs all md5
2000-06-18 23:24:54 +02:00
</programlisting>
</example>
</sect1>
2008-10-23 15:31:10 +02:00
<sect1 id="auth-username-maps">
<title>Username maps</title>
<indexterm zone="auth-username-maps">
<primary>Username maps</primary>
</indexterm>
<para>
When using an external authentication system like Ident or GSSAPI,
the name of the operating system user that initiated the connection may
not be the same as the database user he is requesting to connect as.
In this case, a user name map can be applied to map the operating system
username to a database user, using the <filename>pg_ident.conf</filename>
file. In order to use username mapping, specify
<literal>map</literal>=<replaceable>map-name</replaceable>
in the options field in <filename>pg_hba.conf</filename>. This option is
supported for all authentication methods that receive external usernames.
Since the <filename>pg_ident.conf</filename> file can contain multiple maps,
the name of the map to be used is specified in the
<replaceable>map-name</replaceable> parameter in <filename>pg_hba.conf</filename>
to indicate which map to use for each individual connection.
</para>
<para>
Ident maps are defined in the ident map file, which by default is named
<filename>pg_ident.conf</><indexterm><primary>pg_ident.conf</primary></indexterm>
and is stored in the
cluster's data directory. (It is possible to place the map file
elsewhere, however; see the <xref linkend="guc-ident-file">
configuration parameter.)
The ident map file contains lines of the general form:
<synopsis>
<replaceable>map-name</> <replaceable>system-username</> <replaceable>database-username</>
</synopsis>
Comments and whitespace are handled in the same way as in
<filename>pg_hba.conf</>. The
<replaceable>map-name</> is an arbitrary name that will be used to
refer to this mapping in <filename>pg_hba.conf</filename>. The other
two fields specify which operating system user is allowed to connect
as which database user. The same <replaceable>map-name</> can be
used repeatedly to specify more user-mappings within a single map.
There is no restriction regarding how many database users a given
operating system user can correspond to, nor vice versa.
</para>
<para>
The <filename>pg_ident.conf</filename> file is read on start-up and
when the main server process receives a
<systemitem>SIGHUP</systemitem><indexterm><primary>SIGHUP</primary></indexterm>
signal. If you edit the file on an
active system, you will need to signal the server
(using <literal>pg_ctl reload</> or <literal>kill -HUP</>) to make it
re-read the file.
</para>
<para>
A <filename>pg_ident.conf</filename> file that could be used in
conjunction with the <filename>pg_hba.conf</> file in <xref
linkend="example-pg-hba.conf"> is shown in <xref
linkend="example-pg-ident.conf">. In this example setup, anyone
logged in to a machine on the 192.168 network that does not have the
Unix user name <literal>bryanh</>, <literal>ann</>, or
<literal>robert</> would not be granted access. Unix user
<literal>robert</> would only be allowed access when he tries to
connect as <productname>PostgreSQL</> user <literal>bob</>, not
as <literal>robert</> or anyone else. <literal>ann</> would
only be allowed to connect as <literal>ann</>. User
<literal>bryanh</> would be allowed to connect as either
<literal>bryanh</> himself or as <literal>guest1</>.
</para>
<example id="example-pg-ident.conf">
<title>An example <filename>pg_ident.conf</> file</title>
<programlisting>
# MAPNAME IDENT-USERNAME PG-USERNAME
omicron bryanh bryanh
omicron ann ann
# bob has user name robert on these machines
omicron robert bob
# bryanh can also connect as guest1
omicron bryanh guest1
</programlisting>
</example>
</sect1>
2000-06-18 23:24:54 +02:00
<sect1 id="auth-methods">
<title>Authentication methods</title>
<para>
2004-12-27 00:06:56 +01:00
The following subsections describe the authentication methods in more detail.
2000-06-18 23:24:54 +02:00
</para>
2001-12-27 22:37:34 +01:00
<sect2 id="auth-trust">
<title>Trust authentication</title>
<para>
When <literal>trust</> authentication is specified,
<productname>PostgreSQL</productname> assumes that anyone who can
2004-01-26 06:35:15 +01:00
connect to the server is authorized to access the database with
2005-08-15 01:35:38 +02:00
whatever database user name they specify (including superusers).
2004-12-27 00:06:56 +01:00
Of course, restrictions made in the <literal>database</> and
<literal>user</> columns still apply.
This method should only be used when there is adequate
operating-system-level protection on connections to the server.
2001-12-27 22:37:34 +01:00
</para>
<para>
2003-01-19 01:13:31 +01:00
<literal>trust</> authentication is appropriate and very
convenient for local connections on a single-user workstation. It
is usually <emphasis>not</> appropriate by itself on a multiuser
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
machine. However, you might be able to use <literal>trust</> even
2003-03-13 02:30:29 +01:00
on a multiuser machine, if you restrict access to the server's
Unix-domain socket file using file-system permissions. To do this, set the
2001-12-27 22:37:34 +01:00
<varname>unix_socket_permissions</varname> (and possibly
2003-01-19 01:13:31 +01:00
<varname>unix_socket_group</varname>) configuration parameters as
2003-06-25 03:14:05 +02:00
described in <xref linkend="runtime-config-connection">. Or you
2003-01-19 01:13:31 +01:00
could set the <varname>unix_socket_directory</varname>
configuration parameter to place the socket file in a suitably
restricted directory.
2001-12-27 22:37:34 +01:00
</para>
<para>
2002-01-20 23:19:57 +01:00
Setting file-system permissions only helps for Unix-socket connections.
2003-03-13 02:30:29 +01:00
Local TCP/IP connections are not restricted by it; therefore, if you want
to use file-system permissions for local security, remove the <literal>host ...
2001-12-27 22:37:34 +01:00
127.0.0.1 ...</> line from <filename>pg_hba.conf</>, or change it to a
non-<literal>trust</> authentication method.
</para>
<para>
2003-03-13 02:30:29 +01:00
<literal>trust</> authentication is only suitable for TCP/IP connections
2001-12-27 22:37:34 +01:00
if you trust every user on every machine that is allowed to connect
2002-11-11 21:14:04 +01:00
to the server by the <filename>pg_hba.conf</> lines that specify
2001-12-27 22:37:34 +01:00
<literal>trust</>. It is seldom reasonable to use <literal>trust</>
2003-03-13 02:30:29 +01:00
for any TCP/IP connections other than those from <systemitem>localhost</> (127.0.0.1).
2001-12-27 22:37:34 +01:00
</para>
</sect2>
2001-11-12 20:19:39 +01:00
<sect2 id="auth-password">
2000-06-18 23:24:54 +02:00
<title>Password authentication</title>
2001-05-13 00:51:36 +02:00
<indexterm>
2002-04-04 06:25:54 +02:00
<primary>MD5</>
2001-05-13 00:51:36 +02:00
</indexterm>
2002-04-04 06:25:54 +02:00
<indexterm>
<primary>password</primary>
2003-08-31 19:32:24 +02:00
<secondary>authentication</secondary>
2001-11-12 20:19:39 +01:00
</indexterm>
2001-05-13 00:51:36 +02:00
2001-12-27 22:37:34 +01:00
<para>
2003-03-13 02:30:29 +01:00
The password-based authentication methods are <literal>md5</>,
2008-10-28 13:10:44 +01:00
and <literal>password</>. These methods operate
2001-12-27 22:37:34 +01:00
similarly except for the way that the password is sent across the
2008-10-28 13:10:44 +01:00
connection: respectively, MD5-hashed and clear-text.
2004-12-27 00:06:56 +01:00
</para>
<para>
If you are at all concerned about password
2008-10-28 13:10:44 +01:00
<quote>sniffing</> attacks then <literal>md5</> is preferred.
Plain <literal>password</> should always be avoided if possible.
2001-12-27 22:37:34 +01:00
</para>
2000-06-18 23:24:54 +02:00
<para>
2002-04-04 06:25:54 +02:00
<productname>PostgreSQL</productname> database passwords are
2002-11-11 21:14:04 +01:00
separate from operating system user passwords. The password for
2005-08-15 01:35:38 +02:00
each database user is stored in the <literal>pg_authid</> system
catalog. Passwords can be managed with the SQL commands
2004-12-27 00:06:56 +01:00
<xref linkend="sql-createuser" endterm="sql-createuser-title"> and
<xref linkend="sql-alteruser" endterm="sql-alteruser-title">,
e.g., <userinput>CREATE USER foo WITH PASSWORD 'secret';</userinput>.
By default, that is, if no password has been set up, the stored password
is null and password authentication will always fail for that user.
2000-07-04 18:32:01 +02:00
</para>
2000-06-18 23:24:54 +02:00
</sect2>
2007-07-18 14:00:47 +02:00
<sect2 id="gssapi-auth">
<title>GSSAPI authentication</title>
<indexterm zone="gssapi-auth">
<primary>GSSAPI</primary>
</indexterm>
<para>
<productname>GSSAPI</productname> is an industry-standard protocol
2008-01-05 14:17:00 +01:00
for secure authentication defined in RFC 2743.
2007-07-18 14:00:47 +02:00
<productname>PostgreSQL</productname> supports
<productname>GSSAPI</productname> with <productname>Kerberos</productname>
2008-01-05 14:17:00 +01:00
authentication according to RFC 1964. <productname>GSSAPI</productname>
2007-07-18 14:00:47 +02:00
provides automatic authentication (single sign-on) for systems
that support it. The authentication itself is secure, but the
data sent over the connection will be in clear unless
<acronym>SSL</acronym> is used.
</para>
<para>
When <productname>GSSAPI</productname> uses
<productname>Kerberos</productname>, it uses a standard principal
in format
<literal><replaceable>servicename</>/<replaceable>hostname</>@<replaceable>realm</></literal>. For information about the parts of the principal, and
how to set up the required keys, see <xref linkend="kerberos-auth">.
GSSAPI support has to be enabled when <productname>PostgreSQL</> is built;
see <xref linkend="installation"> for more information.
</para>
2008-10-23 15:31:10 +02:00
<para>
The following configuration options are supported for <productname>GSSAPI</productname>:
<variablelist>
<varlistentry>
<term>map</term>
<listitem>
<para>
Allows for mapping between system and database usernames. See
<xref linkend="auth-username-maps"> for details.
</para>
</listitem>
</varlistentry>
</variablelist>
</para>
2007-07-18 14:00:47 +02:00
</sect2>
2007-07-23 12:16:54 +02:00
<sect2 id="sspi-auth">
<title>SSPI authentication</title>
<indexterm zone="sspi-auth">
<primary>SSPI</primary>
</indexterm>
<para>
<productname>SSPI</productname> is a <productname>Windows</productname>
technology for secure authentication with single sign-on.
<productname>PostgreSQL</productname> will use SSPI in
<literal>negotiate</literal> mode, which will use
<productname>Kerberos</productname> when possible and automatically
fall back to <productname>NTLM</productname> in other cases.
<productname>SSPI</productname> authentication only works when both
server and client are running <productname>Windows</productname>.
</para>
<para>
When using <productname>Kerberos</productname> authentication,
<productname>SSPI</productname> works the same way
<productname>GSSAPI</productname> does. See <xref linkend="gssapi-auth">
for details.
</para>
2008-10-23 15:31:10 +02:00
<para>
The following configuration options are supported for <productname>SSPI</productname>:
<variablelist>
<varlistentry>
<term>map</term>
<listitem>
<para>
Allows for mapping between system and database usernames. See
<xref linkend="auth-username-maps"> for details.
</para>
</listitem>
</varlistentry>
</variablelist>
</para>
2007-07-23 12:16:54 +02:00
</sect2>
2000-08-25 12:00:35 +02:00
<sect2 id="kerberos-auth">
2000-06-18 23:24:54 +02:00
<title>Kerberos authentication</title>
2001-05-13 00:51:36 +02:00
<indexterm zone="kerberos-auth">
<primary>Kerberos</primary>
</indexterm>
2007-11-14 15:25:55 +01:00
<note>
<para>
Native Kerberos authentication has been deprecated and should be used
only for backward compatibility. New and upgraded installations are
encouraged to use the industry-standard <productname>GSSAPI</productname>
authentication (see <xref linkend="gssapi-auth">) instead.
</para>
</note>
2000-06-18 23:24:54 +02:00
<para>
<productname>Kerberos</productname> is an industry-standard secure
2003-08-17 01:33:49 +02:00
authentication system suitable for distributed computing over a public
network. A description of the <productname>Kerberos</productname> system
2004-12-27 00:06:56 +01:00
is far beyond the scope of this document; in full generality it can be
2005-04-09 05:52:43 +02:00
quite complex (yet powerful). The
<ulink url="http://www.nrl.navy.mil/CCS/people/kenh/kerberos-faq.html">
Kerberos <acronym>FAQ</></ulink> or
2005-06-04 22:42:43 +02:00
<ulink url="http://web.mit.edu/kerberos/www/">MIT Kerberos page</ulink>
2005-08-15 01:35:38 +02:00
can be good starting points for exploration.
2003-08-17 01:33:49 +02:00
Several sources for <productname>Kerberos</> distributions exist.
2006-11-23 06:39:17 +01:00
<productname>Kerberos</productname> provides secure authentication but
does not encrypt queries or data passed over the network; for that
use <acronym>SSL</acronym>.
2003-08-17 01:33:49 +02:00
</para>
<para>
2005-08-15 01:35:38 +02:00
<productname>PostgreSQL</> supports Kerberos version 5. Kerberos
support has to be enabled when <productname>PostgreSQL</> is built;
see <xref linkend="installation"> for more information.
2000-06-18 23:24:54 +02:00
</para>
<para>
2001-11-21 06:53:41 +01:00
<productname>PostgreSQL</> operates like a normal Kerberos service.
2001-11-19 04:58:25 +01:00
The name of the service principal is
2005-06-04 22:42:43 +02:00
<literal><replaceable>servicename</>/<replaceable>hostname</>@<replaceable>realm</></literal>.
2005-08-15 01:35:38 +02:00
</para>
<para>
2005-06-04 22:42:43 +02:00
<replaceable>servicename</> can be set on the server side using the
<xref linkend="guc-krb-srvname"> configuration parameter, and on the
2005-08-15 01:35:38 +02:00
client side using the <literal>krbsrvname</> connection parameter. (See
also <xref linkend="libpq-connect">.) The installation default can be
changed from the default <literal>postgres</literal> at build time using
<literal>./configure --with-krb-srvnam=whatever</>. In most environments,
this parameter never needs to be changed. However, to support multiple
<productname>PostgreSQL</> installations on the same host it is necessary.
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
Some Kerberos implementations might also require a different service name,
2005-08-15 01:35:38 +02:00
such as Microsoft Active Directory which requires the service name
to be in uppercase (<literal>POSTGRES</literal>).
</para>
<para>
2003-03-13 02:30:29 +01:00
<replaceable>hostname</> is the fully qualified host name of the
2002-04-04 06:25:54 +02:00
server machine. The service principal's realm is the preferred realm
of the server machine.
2001-11-19 04:58:25 +01:00
</para>
<para>
2005-08-15 01:35:38 +02:00
Client principals must have their <productname>PostgreSQL</> database user
2002-04-04 06:25:54 +02:00
name as their first component, for example
2007-11-09 18:31:07 +01:00
<literal>pgusername@realm</>. By default, the realm of the client is
2007-11-10 00:52:08 +01:00
not checked by <productname>PostgreSQL</>. If you have cross-realm
authentication enabled and need to verify the realm, use the
<xref linkend="guc-krb-realm"> parameter.
2001-11-19 04:58:25 +01:00
</para>
<para>
2005-06-04 22:42:43 +02:00
Make sure that your server keytab file is readable (and preferably
2004-03-09 17:57:47 +01:00
only readable) by the <productname>PostgreSQL</productname> server
2004-12-27 00:06:56 +01:00
account. (See also <xref linkend="postgres-user">.) The location
2004-03-09 17:57:47 +01:00
of the key file is specified by the <xref
2004-12-27 00:06:56 +01:00
linkend="guc-krb-server-keyfile"> configuration
2005-06-27 04:04:26 +02:00
parameter. The default is
2004-03-09 17:57:47 +01:00
<filename>/usr/local/pgsql/etc/krb5.keytab</> (or whichever
2005-06-27 04:04:26 +02:00
directory was specified as <varname>sysconfdir</> at build time).
2000-06-18 23:24:54 +02:00
</para>
2000-07-15 23:35:47 +02:00
<para>
2005-08-15 01:35:38 +02:00
The keytab file is generated by the Kerberos software; see the
Kerberos documentation for details. The following example is
for MIT-compatible Kerberos 5 implementations:
2000-07-15 23:35:47 +02:00
<screen>
2000-09-06 21:54:52 +02:00
<prompt>kadmin% </><userinput>ank -randkey postgres/server.my.domain.org</>
<prompt>kadmin% </><userinput>ktadd -k krb5.keytab postgres/server.my.domain.org</>
2000-07-15 23:35:47 +02:00
</screen>
</para>
<para>
2001-11-19 04:58:25 +01:00
When connecting to the database make sure you have a ticket for a
2005-08-15 01:35:38 +02:00
principal matching the requested database user name. For example, for
2002-04-04 06:25:54 +02:00
database user name <literal>fred</>, both principal
2001-11-19 04:58:25 +01:00
<literal>fred@EXAMPLE.COM</> and
2005-08-15 01:35:38 +02:00
<literal>fred/users.example.com@EXAMPLE.COM</> could be used to
2002-04-04 06:25:54 +02:00
authenticate to the database server.
2000-07-15 23:35:47 +02:00
</para>
<para>
2005-04-09 05:52:43 +02:00
If you use <ulink url="http://modauthkerb.sf.net">
<application>mod_auth_kerb</application></ulink>
2003-06-06 20:09:17 +02:00
and <application>mod_perl</application> on your
2002-04-04 06:25:54 +02:00
<productname>Apache</productname> web server, you can use
<literal>AuthType KerberosV5SaveCredentials</literal> with a
<application>mod_perl</application> script. This gives secure
database access over the web, no extra passwords required.
2000-07-15 23:35:47 +02:00
</para>
2000-06-18 23:24:54 +02:00
</sect2>
2002-09-14 20:35:46 +02:00
<sect2 id="auth-ident">
2000-06-18 23:24:54 +02:00
<title>Ident-based authentication</title>
2001-05-13 00:51:36 +02:00
<indexterm>
<primary>ident</primary>
</indexterm>
2002-09-14 20:35:46 +02:00
<para>
2004-12-27 00:06:56 +01:00
The ident authentication method works by obtaining the client's
2008-09-15 14:32:57 +02:00
operating system user name, then optionally determining the allowed
database user names using a map file that lists the permitted
2004-12-27 00:06:56 +01:00
corresponding pairs of names. The determination of the client's
2002-09-14 20:35:46 +02:00
user name is the security-critical point, and it works differently
depending on the connection type.
</para>
2008-10-23 15:31:10 +02:00
<para>
The following configuration options are supported for <productname>GSSAPI</productname>:
<variablelist>
<varlistentry>
<term>map</term>
<listitem>
<para>
Allows for mapping between system and database usernames. See
<xref linkend="auth-username-maps"> for details.
</para>
</listitem>
</varlistentry>
</variablelist>
</para>
2002-09-14 20:35:46 +02:00
<sect3>
<title>Ident Authentication over TCP/IP</title>
2000-06-18 23:24:54 +02:00
<para>
The <quote>Identification Protocol</quote> is described in
2008-01-05 14:17:00 +01:00
RFC 1413. Virtually every Unix-like
2001-03-15 21:01:32 +01:00
operating system ships with an ident server that listens on TCP
2000-07-04 18:32:01 +02:00
port 113 by default. The basic functionality of an ident server
2000-06-18 23:24:54 +02:00
is to answer questions like <quote>What user initiated the
connection that goes out of your port <replaceable>X</replaceable>
and connects to my port <replaceable>Y</replaceable>?</quote>.
2001-11-21 06:53:41 +01:00
Since <productname>PostgreSQL</> knows both <replaceable>X</> and
2000-07-04 18:32:01 +02:00
<replaceable>Y</> when a physical connection is established, it
can interrogate the ident server on the host of the connecting
client and could theoretically determine the operating system user
for any given connection this way.
2000-06-18 23:24:54 +02:00
</para>
<para>
The drawback of this procedure is that it depends on the integrity
of the client: if the client machine is untrusted or compromised
an attacker could run just about any program on port 113 and
return any user name he chooses. This authentication method is
therefore only appropriate for closed networks where each client
machine is under tight control and where the database and system
2001-10-05 00:27:18 +02:00
administrators operate in close contact. In other words, you must
trust the machine running the ident server.
Heed the warning:
2000-06-18 23:24:54 +02:00
<blockquote>
<attribution>RFC 1413</attribution>
<para>
The Identification Protocol is not intended as an authorization
2001-07-11 23:27:07 +02:00
or access control protocol.
2000-06-18 23:24:54 +02:00
</para>
</blockquote>
</para>
2005-01-28 23:38:37 +01:00
<para>
Some ident servers have a nonstandard option that causes the returned
user name to be encrypted, using a key that only the originating
machine's administrator knows. This option <emphasis>must not</> be
used when using the ident server with <productname>PostgreSQL</>,
since <productname>PostgreSQL</> does not have any way to decrypt the
returned string to determine the actual user name.
</para>
2002-09-14 20:35:46 +02:00
</sect3>
<sect3>
<title>Ident Authentication over Local Sockets</title>
2000-06-18 23:24:54 +02:00
2001-08-02 01:25:39 +02:00
<para>
2002-04-04 06:25:54 +02:00
On systems supporting <symbol>SO_PEERCRED</symbol> requests for
2002-09-14 20:35:46 +02:00
Unix-domain sockets (currently <systemitem
class="osname">Linux</>, <systemitem class="osname">FreeBSD</>,
2008-09-15 14:41:54 +02:00
<systemitem class="osname">NetBSD</>, <systemitem class="osname">OpenBSD</>,
2008-11-18 14:10:20 +01:00
<systemitem class="osname">BSD/OS</>, and <systemitem class="osname">Solaris</systemitem>), ident authentication can also
2003-02-13 06:47:46 +01:00
be applied to local connections. In this case, no security risk is added by
2002-09-14 20:35:46 +02:00
using ident authentication; indeed it is a preferable choice for
local connections on such systems.
2001-08-02 01:25:39 +02:00
</para>
2002-09-14 20:35:46 +02:00
<para>
On systems without <symbol>SO_PEERCRED</> requests, ident
authentication is only available for TCP/IP connections. As a
2004-12-27 00:06:56 +01:00
work-around, it is possible to specify the <systemitem
2002-09-14 20:35:46 +02:00
class="systemname">localhost</> address <systemitem
class="systemname">127.0.0.1</> and make connections to this
2004-12-27 00:06:56 +01:00
address. This method is trustworthy to the extent that you trust
the local ident server.
2002-09-14 20:35:46 +02:00
</para>
</sect3>
</sect2>
2006-06-16 17:16:16 +02:00
<sect2 id="auth-ldap">
<title>LDAP authentication</title>
<indexterm zone="auth-ldap">
<primary>LDAP</primary>
</indexterm>
<para>
This authentication method operates similarly to
<literal>password</literal> except that it uses LDAP
2008-10-23 15:31:10 +02:00
as the password verification method. LDAP is used only to validate
2006-06-16 17:16:16 +02:00
the user name/password pairs. Therefore the user must already
exist in the database before LDAP can be used for
2008-10-23 15:31:10 +02:00
authentication.
2006-06-16 17:16:16 +02:00
</para>
2008-10-23 15:31:10 +02:00
2006-06-16 17:16:16 +02:00
<para>
2008-10-23 15:31:10 +02:00
The server will bind to the distinguished name constructed as
<replaceable>prefix</> <replaceable>username</> <replaceable>suffix</>.
2006-06-16 17:16:16 +02:00
before the bind. Typically, the prefix parameter is used to specify
<replaceable>cn=</>, or <replaceable>DOMAIN\</> in an Active
2008-10-23 15:31:10 +02:00
Directory environment, and suffix is used to specify the remaining part
of the DN in a non-Active Directory environment.
2006-06-16 17:16:16 +02:00
</para>
2008-10-23 15:31:10 +02:00
<para>
The following configuration options are supported for LDAP:
<variablelist>
<varlistentry>
<term>ldapserver</term>
<listitem>
<para>
Name or IP of LDAP server to connect to.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldapprefix</term>
<listitem>
<para>
String to prepend to the username when building the base DN to
bind as.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldapsuffix</term>
<listitem>
<para>
String to append to the username when building the base DN to
bind as.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldapport</term>
<listitem>
<para>
Port number on LDAP server to connect to. If no port is specified,
the default port in the LDAP library will be used.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldaptls</term>
<listitem>
<para>
Set to 1 to make the connection between PostgreSQL and the
LDAP server use TLS encryption. Note that this only encrypts
the traffic to the LDAP server - the connection to the client
may still be unencrypted unless TLS is used there as well.
</para>
</listitem>
</varlistentry>
</variablelist>
</para>
<note>
<para>
Since LDAP often uses commas and spaces to separate the different
parts of a DN, it is advised to always use double-quoted parameter
values when configuring LDAP options, such as:
</para>
</note>
<synopsis>
ldapserver=ldap.example.net prefix="cn=" suffix="dc=example, dc=net"
</synopsis>
2006-06-16 17:16:16 +02:00
</sect2>
2008-11-20 12:48:26 +01:00
<sect2 id="auth-cert">
<title>Certificate authentication</title>
<indexterm zone="auth-cert">
<primary>Certificate</primary>
</indexterm>
<para>
This authentication method uses SSL client certificates to perform
authentication. It is therefore only available for SSL connections.
When using this authentication method, the server will require that
the client provide a certificate. No password prompt will be sent
to the client. The <literal>cn</literal> attribute of the certificate
will be matched with the username the user is trying to log in as,
and if they match the login will be allowed. Username mapping can be
used if the usernames don't match.
</para>
</sect2>
2002-09-14 20:35:46 +02:00
<sect2 id="auth-pam">
2004-12-27 00:06:56 +01:00
<title>PAM authentication</title>
2002-09-14 20:35:46 +02:00
2003-08-31 19:32:24 +02:00
<indexterm zone="auth-pam">
<primary>PAM</primary>
</indexterm>
2002-09-14 20:35:46 +02:00
<para>
2003-03-13 02:30:29 +01:00
This authentication method operates similarly to
<literal>password</literal> except that it uses PAM (Pluggable
2002-09-14 20:35:46 +02:00
Authentication Modules) as the authentication mechanism. The
2008-10-23 15:31:10 +02:00
default PAM service name is <literal>postgresql</literal>.
2005-11-05 00:14:02 +01:00
PAM is used only to validate user name/password pairs.
2005-04-27 22:11:07 +02:00
Therefore the user must already exist in the database before PAM
2005-04-26 05:01:09 +02:00
can be used for authentication. For more information about
PAM, please read the <ulink url="http://www.kernel.org/pub/linux/libs/pam/">
2005-04-09 05:52:43 +02:00
<productname>Linux-PAM</> Page</ulink>
and the <ulink url="http://www.sun.com/software/solaris/pam/">
<systemitem class="osname">Solaris</> PAM Page</ulink>.
2002-09-14 20:35:46 +02:00
</para>
2007-09-14 05:53:54 +02:00
2008-10-23 15:31:10 +02:00
<para>
The following configuration options are supported for PAM:
<variablelist>
<varlistentry>
<term>pamservice</term>
<listitem>
<para>
PAM service name.
</para>
</listitem>
</varlistentry>
</variablelist>
</para>
2007-09-14 05:53:54 +02:00
<note>
<para>
2007-12-29 05:15:38 +01:00
If PAM is set up to read <filename>/etc/shadow</>, authentication
will fail because the PostgreSQL server is started by a non-root
user. However, this is not an issue with LDAP or other authentication
methods.
2007-09-14 05:53:54 +02:00
</para>
</note>
2000-06-18 23:24:54 +02:00
</sect2>
</sect1>
<sect1 id="client-authentication-problems">
<title>Authentication problems</title>
<para>
Genuine authentication failures and related problems generally
manifest themselves through error messages like the following.
</para>
<para>
2005-01-23 01:30:59 +01:00
<programlisting>
2003-09-13 00:17:24 +02:00
FATAL: no pg_hba.conf entry for host "123.123.123.123", user "andym", database "testdb"
2005-01-23 01:30:59 +01:00
</programlisting>
2002-04-04 06:25:54 +02:00
This is what you are most likely to get if you succeed in contacting
the server, but it does not want to talk to you. As the message
suggests, the server refused the connection request because it found
2005-08-15 01:35:38 +02:00
no matching entry in its <filename>pg_hba.conf</filename>
2000-06-18 23:24:54 +02:00
configuration file.
</para>
<para>
2005-01-23 01:30:59 +01:00
<programlisting>
2003-09-13 00:17:24 +02:00
FATAL: Password authentication failed for user "andym"
2005-01-23 01:30:59 +01:00
</programlisting>
2002-04-04 06:25:54 +02:00
Messages like this indicate that you contacted the server, and it is
willing to talk to you, but not until you pass the authorization
method specified in the <filename>pg_hba.conf</filename> file. Check
the password you are providing, or check your Kerberos or ident
software if the complaint mentions one of those authentication
types.
2000-06-18 23:24:54 +02:00
</para>
<para>
2005-01-23 01:30:59 +01:00
<programlisting>
2003-09-13 00:17:24 +02:00
FATAL: user "andym" does not exist
2005-01-23 01:30:59 +01:00
</programlisting>
2001-11-28 21:49:10 +01:00
The indicated user name was not found.
2000-06-18 23:24:54 +02:00
</para>
<para>
2005-01-23 01:30:59 +01:00
<programlisting>
2003-09-13 00:17:24 +02:00
FATAL: database "testdb" does not exist
2005-01-23 01:30:59 +01:00
</programlisting>
2001-11-28 21:49:10 +01:00
The database you are trying to connect to does not exist. Note that
if you do not specify a database name, it defaults to the database
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
user name, which might or might not be the right thing.
2000-06-18 23:24:54 +02:00
</para>
2001-08-02 01:25:39 +02:00
2003-09-13 00:17:24 +02:00
<tip>
2001-08-02 01:25:39 +02:00
<para>
Update documentation on may/can/might:
Standard English uses "may", "can", and "might" in different ways:
may - permission, "You may borrow my rake."
can - ability, "I can lift that log."
might - possibility, "It might rain today."
Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice. Similarly, "It may crash" is better stated, "It might crash".
Also update two error messages mentioned in the documenation to match.
2007-01-31 21:56:20 +01:00
The server log might contain more information about an
2002-04-04 06:25:54 +02:00
authentication failure than is reported to the client. If you are
confused about the reason for a failure, check the log.
2001-08-02 01:25:39 +02:00
</para>
2003-09-13 00:17:24 +02:00
</tip>
2000-06-18 23:24:54 +02:00
</sect1>
</chapter>