Chapter 8. Authentication and Encryption

Table of Contents
Client Authentication
Encrypting sessions

This chapter documents the fundamental concepts involved with authenticating and encrypting a client session to the PostgreSQL server. This includes how to correctly configure the pg_hba.conf file for a variety of authentication schemes, as well as a few common ways to encrypt your client connections.

Client Authentication

Client authentication is a central feature to PostgreSQL. Without it, you would either have to sacrifice remote connectivity, or blindly allow anyone to connect to your database and retrieve, or even modify your data. PostgreSQL has several different types of client authentication at its disposal. As the site administrator, you need to decide which one is best for your system.

As of PostgreSQL 7.1.x, host-based client access is specified in the pg_hba.conf file. The rights and restrictions described in this file should not be confused with a PostgreSQL user's rights to objects within the database. The pg_hba.conf file allows you to set the type of host-based authentication to be used. This authentication is performed before PostgreSQL establishes a connection to the intended database, where user rights would be relevant.

Note: The pg_hba.conf is located in the PostgreSQL data directory (e.g., /usr/local/pgsql/data ), and is installed automatically upon the execution of the initdb command when PostgreSQL is installed.

PostgreSQL's host-based authentication is flexible, featuring a wide variety of configurable options. You may restrict database access to specific hosts, as well as allow access to a range of IP addresses by using netmasks. Each configured host has its own host record, which is a single line in the pg_hba.conf file.

With these host records, you may specify access either to a particular database or all databases. Furthermore, you may require a user from a specified host to authenticate via the PostgreSQL users table after qualifying for a connection.

Put simply, the pg_hba.conf file allows you to determine who is allowed to connect to which databases from what machines, and to what degree they must prove their authenticity to gain access.

Warning

Through remote password-based authentication, passwords may be transmitted in clear text depending on whether or not you are using encrypted sessions. Be sure that you understand how your application is communicating with PostgreSQL before allowing users to remotely connect to a PostgreSQL database.

Password Authentication

Passwords allow PostgreSQL users a way to identify themselves and prevent unauthorized individuals from connecting with a user that is not theirs. As of PostgreSQL 7.1.x, user passwords are stored in plain text in the pg_shadow system table. The structure of this table is illustrated in Table 8-1. Note that while the passwords are stored as plain text, only PostgreSQL superusers are allowed to view the pg_shadow table.

Table 8-1. The pg_shadow table

Column

Type

usename

name

usesysid

integer

usecreatedb

boolean

usetrace

boolean

usesuper

boolean

usecatupd

boolean

passwd

text

valuntil

abstime

The pg_shadow table is a system table, and thus is accessible from any database. It follows, therefore, that users are not assigned to a specific database. If a user exists in the pg_shadow table, that user will be able to connect to any database on the server machine, though not necessarily from any remote machine (depending on your configuration).

Users typically set passwords in PostgreSQL when the user is created (with the CREATE USER command) or after the user has been created (using the ALTER USER command). Alternatively, you may manually modify a user's password by using an UPDATE statement. (For a more detailed explanation about defining passwords for users, see Chapter 10.)

If a password is not set, a user's password defaults to NULL. If password-based authentication is enabled in the pg_hba.conf file, connection attempts will always fail for such a user. Conversely, if the host that establishes the connection is a trusted host (such as localhost, by default), anyone from the trusted host may connect as a user with a NULL password. In fact, passwords are ignored entirely for trusted hosts.

Note: The GRANT command allows you to restrict or allow a variety of access types to tables within a database. See Chapter 10 for more on this topic.

Unless your needs for security are very minimal, you will not want to rely on password-only authentication with your PostgreSQL server. Using a password-only method to authenticate users will allow any verified user access to any database on the system, and authenticating with a password over clear text can result in unauthorized individuals acquiring user passwords. If you are likely to have your database connected to the Internet in some fashion, we strongly suggest that you read the following sections. These cover the use of the pg_hba.conf file and session encryption.

The pg_hba.conf file

We mentioned earlier in this section that the pg_hba.conf file enables client authentication between the PostgreSQL server and the client application. This file consists of a series of entries, which define a host and its associated permissions (e.g., the database it is allowed to connect to, the authentication method to use, and so on).

When an application requests a connection, the request will specify a PostgreSQL username and database with which it intends to connect to PostgreSQL. Optionally, a password may be provided, depending on the expected configuration for the connecting host.

Note: PostgreSQL has its own user and password tables, which are separate from system accounts. It is not required that your PostgreSQL users match users available to the operating system.

When PostgreSQL receives a connection request it will check the pg_hba.conf file to verify that the machine from which the application is requesting a connection has rights to connect to the specified database. If the machine requesting access has permission to connect, PostgreSQL will check the conditions that the application must meet in order to successfully authenticate. This affects connections that are initiated locally as well as remotely.

PostgreSQL will check the authentication method via the pg_hba.conf for every connection request. This check is performed every time a new connection is requested from the PostgreSQL server, so there is no need to re-start PostgreSQL after you add, modify or remove an entry in the pg_hba.conf file. Example 8-1 is a simple example of the pg_hba.conf file.

Example 8-1. A simple pg_hba.conf file

#                   PostgreSQL HOST ACCESS CONTROL FILE
#

local  all                                          trust
host   all        127.0.0.1      255.255.255.255    trust
host   booktown   192.168.1.3    255.255.255.255    ident    sales
host   all        192.168.1.4    255.255.255.255    ident    audit

When a connection is initialized, PostgreSQL will read through the pg_hba.conf one entry at a time, from the top down. As soon a matching record is found, PostgreSQL will stop searching and allow or reject the connection, based on the found entry. If PostgreSQL does not find a matching entry in the pg_hba.conf file, the connection fails completely.

Table-level permissions still apply to a database, even if a user has permissions to connect to the database. If you can connect, but cannot select data from a table, you may want to verify that your connected user has permission to use SELECT on that table. Using the psql command-line application, you can check the permissions of the tables within a database by using the \z slash command. From any other interface to PostgreSQL, use the query demonstrated in Example 8-2 to see the same information provided by the \z slash command.

Example 8-2. Checking user permissions

testdb=# SELECT relname as "Relation", relacl as "Access permissions"
testdb-#        FROM pg_class
testdb-#        WHERE  relkind IN ('r', 'v', 'S')
testdb-#        AND relname !~ '^pg_'
testdb-#        ORDER BY relname;
 Relation |     Access permissions
----------+----------------------------------
 foo      | {"=arwR","jdrake=arwR"}
 my_list  | {"=","jdrake=arwR","jworsley=r"}
(2 rows)

Structure of the pg_hba.conf file

The pg_hba.conf file contains sequential entries that define the settings PostgreSQL should use during the client authentication process for a specified host. This file is designed to be easily customizable to your system needs.

Within this file, you may associate a TCP/IP host address (or a range of addresses) with a particular database (or all databases), and one of several available authentication methods. You may also specify access for local connections using the term localhost, or 127.0.0.1, rather than using the system's external IP address. Several syntax rules apply to the pg_hba.conf.

First, you may only place one host record per line in the file. Subsequently, host records are not allowed to wrap across multiple lines. Second, each host record must contain multiple fields, which must be separated by either tabs or spaces. The number of fields in a host record is directly related to the type of host entry being defined. Example 8-3 shows two host records, the first with the fields separated by spaces, and the second with the file separated by tabs.

Example 8-3. A valid pg_hba.conf entry with spaces and tabs

host all 127.0.0.1 255.255.255.255 trust
host	all	127.0.0.1	255.255.255.255	trust

Commenting is allowed within the pg_hba.conf by placing a hash mark (#) at the beginning of each line being commented. Example 8-4 demonstrates valid commented lines.

Example 8-4. Valid pg_hba.conf comments

# Book Town host entries
#
#
host all 127.0.0.1 255.255.255.255 trust

Regarding the actual form of each host record, there are three general types available in the pg_hba.conf (the type keyword is always the first word in the host record):

host

A host entry is used to specify remote hosts that are allowed to connect to the PostgreSQL server. PostgreSQL's postmaster backend must be running with the -i option (TCP/IP) in order for a host entry to work correctly.

local

A local entry is semantically the same as a host entry. However, you do not need to specify a host that is allowed to connect. The local entry is used for client connections that are initiated from the same machine that the PostgreSQL server is operating on.

hostssl

A hostssl entry is user to specify hosts (remote or local) that are allowed to connect to the PostgreSQL server using SSL. The use of SSL insures that all communication between the client and the server is encrypted. In order for this to work, both the client and the server must support SSL. The postmaster backend must be running with the -l (SSL) and -i (TCP/IP) options.

Note: See Chapter 9 for more on how to start the postmaster process with the appropriate run-time options.

Example 8-5 illustrates the general syntax for each type of host record available within the pg_hba.conf file. Notice that the format is essentially identical for each record, with the exception that a local record does not require an IP address or netmask to be specified, as the connection is assumed to be from the same machine on which PostgreSQL is running.

Example 8-5. Host entry syntax

# A "local" record.
local	database	auth_method	[ auth_option ]     

# A "host" record.
host	database	ip_addr	netmask	auth_method	[ auth_option ]

# A "hostssl" record.
hostssl	database	ip_addr	netmask	auth_method	[ auth_option ]

Note: Remember that each entry in the pg_hba.conf must be a single line. You cannot word wrap or use line breaks.

The following list is a description of the keywords for the pg_hba.conf entries mentioned previously:

database

This is the database name that the specified host is allowed to connect to. The database keyword has three possible values:

all

The all keyword specifies that the client connecting can connect to any database the PostgreSQL server is hosting.

sameuser

The sameuser keyword specifies that the client can only connect to a database that matches the clients authenticated user name.

name

A specific name may be specified, so that the client can only connect to the database as specified by name.

ip_addr, netmask

The ip_addr and netmask fields specify either a specific IP address, or range of IP addresses, that are allowed to connect to the PostgreSQL server. Such a range can by specified by describing an IP network with an associated netmask. Otherwise, for a single IP address, the netmask field should be set to 255.255.255.255.

If you are unsure of how to specify a netmask, view the online Linux Networking HOWTO, at http://www.thelinuxreview.com/howto/networking, or consult your system administrator.

auth_method

The authentication method specifies the type of authentication the server should use for a user trying to connect to PostgreSQL. The following is a list of options available for auth_method:

trust

The trust method allows any user from the defined host to connect to a PostgreSQL database without the use of a password, as any PostgreSQL user. You are trusting the host-based authentication with the use of this method, and any user on the specified host. This is a dangerous condition if the specified host is not a secure machine, or provides access to users unknown to you.

reject

The reject method automatically denies access to PostgreSQL for that host or user. This can be a prudent setting for sites that you know are never allowed to connect to your database server.

password

The password method specifies that a password must exist for a connecting user. The use of this method will require the connecting user to supply a password that matches the password found in the global pg_shadow system table for their username. If you use the password method, the password will be sent in clear text.

crypt

The crypt method is similar to the password method. When using crypt, the password is not sent in clear text, but through a simple form of encryption. The use of this method is not very secure, but is better than using the clear text password method.

krb4, krb5

The krb4 and krb5 methods are used to specify Version 4 or 5 of the Kerberos authentication system. The installation and configuration of Kerberos is beyond the scope of this book, but if you wish to authenticate via Kerberos, these methods are available.

ident

The ident method specifies that an ident map should be used when a host is requesting connections from a valid IP address listed in the pg_hba.conf file. This method requires one option.

The required option may be either the special term sameuser, or a named map that is defined within the pg_ident.conf file. For more information on defining an ident map, see the Section called The pg_ident.conf file."

auth_option

The auth_option field may or may not be required, based on the type of authentication method that is used; as of PostgreSQL 7.1.x, only the ident method requires an option.

Warning

We do not suggest the use of either password or crypt without the use of an external encryption mechanism. See the Section called Encrypting sessions" in this chapter for information on installing a central encryption mechanism for all of your PostgreSQL traffic.

Example pg_hba.conf entries

This section contains a series of examples that can be used within the pg_hba.conf. To begin, the host record within Example 8-6 allows a single machine with the IP address 192.168.1.10 to connect to any database as any user, without the use of a password. This is because it is configured with the all and trust terms, respectively.

Example 8-6. Single host entry

host	all	192.168.1.10	255.255.255.255	trust

Example 8-7 shows a host record which will reject all users from host 192.168.1.10, for any requested database. This is set by the use of the terms all and reject as the database target and authentication method, respectively.

Example 8-7. Rejection entry

host	all	192.168.1.10	255.255.255.255	reject

The host record in Example 8-8 will allow any user with the IP of 192.168.1.10, and a valid password, to connect to the database template1. The password will be encrypted during authentication because of the use of the term crypt.

Example 8-8. Single host, single database entry

host	template1	192.168.1.10	255.255.255.255	crypt

The host record in Example 8-9 allows a small subnet of computers to access any database, without the need of a password. This subnet describes any IP from 192.168.1.1 to 192.168.1.15. Again, if you are unsure of how to configure your netmask, consult your network administrator, or view the Linux Networking HOWTO at http://www.thelinuxreview.com/howto/networking.

Example 8-9. Small network connection entry

host	all	192.168.1.0	255.255.255.240	trust

Expanding on the use of subnets, the host record in Example 8-10 allows any machine on the 192.168.1 block to connect to the booktown database, without the use of a password.

Example 8-10. Larger network connection entry

host	booktown	192.168.1.0	255.255.255.0	trust

Remember, as stated earlier in this section, each host record line is read in succession from the top of the file to the bottom. The first record which matches the host attempting to connect is used. If no matching record is found, connection is completely disallowed.

The pg_ident.conf file

When specifying the ident term as a host record's authentication method, PostgreSQL uses the pg_ident.conf file to map the identifying username to a PostgreSQL username. The identifying username is the name provided by the connecting client's identd service (RFC 1413), which is required to identify the name of the system account initiating the connection. This method is similar to the trust method, but restricts access based on the identifying username.

As stated in the specification for the ident protocol, "The Identification Protocol is not intended as an authorization or access control protocol." This is only a useful method of identification for secure, controlled machines, and is not intended as a means for secure control from a wide array of external machines. This is because an identd daemon merely returns an arbitrary username describing the current system user. For example, allowing the username jworsley from an entire subnet of IP addresses would create a serious security hole, because anyone with a machine in that subnet could create a user named jworsley and become "authenticated" as a result.

The pg_ident.conf file should be located in the same path as the pg_hba.conf file. This should be the path defined by the PGDATA environment variable (e.g., /usr/local/pgsql/data). Like the pg_hba.conf, changes to the pg_ident.conf file do not require PostgreSQL to be re-started.

The content of the pg_ident.conf associates identifying usernames with PostgreSQL usernames via definitions called ident maps. This is useful for users whose system usernames do not match their PostgreSQL usernames. Some rules you should keep in mind when defining and using an ident map are:

  • Each ident map member is defined on a single line, which associates a map name with an identifying username, and a translated PostgreSQL username.

  • The pg_ident.conf file can contain multiple map names. Each group of single lines with the same associative map name are considered a single map.

  • The pg_hba.conf file determines the types of connections that relate to users in this file.

A single line record to define an ident map consist of 3 tokens: the name of the map, the identifying username, and the translated PostgreSQL username. This syntax is entered as follows, where each token is separated by spaces, or tabs:

mapname identname postgresqlname

mapname

The map name used in the pg_hba.conf file to refer to the ident map.

identname

The identifying username, which is generally the name of the system user attempting to establish a connection to the database. This is the name provided by the identd daemon, which must be running on the system attempting to connect.

postgresqlname

The database username which is allowed for the preceding identifying username. You may specify several lines with the same identname, but with different postgresqlname values, in order to allow a single system user access to several accounts, which do not all need to be on the same database.

As an example, suppose that the Book Town server has a set of system accounts named jdrake, jworsley, and auditor, used for two salespeople and an internal auditor, respectively.

You may wish to create a pair of ident maps for these two groups of users. Suppose that the sales department's workstation has an IP address of 192.168.1.3, and only needs access to the booktown database, while the audit department's workstation has an IP address of 192.168.1.4, and requires access to all databases. This scenario might result in a pga_hba.conf, such as the one displayed in Example 8-11.

Example 8-11. An ident configuration in pg_hba.conf

host   booktown   192.168.1.3    255.255.255.255    ident    sales
host   all        192.168.1.4    255.255.255.255    ident    audit

This host access configuration states that the sales machine may connect to the booktown database using an ident map named sales, and the audit workstation may connect to any database using an ident map named audit. Each of these maps must then be configured within the pg_ident.conf file. Example 8-12 demonstrates such a configuration.

Example 8-12. A pg_ident.conf configuration

# MAP   IDENT        POSTGRESQL_USERNAME
sales   jdrake       sales
sales   jworsley     sales
audit   auditor      sales
audit   auditor      postgres

The file shown in Example 8-12 allows either of the system users jdrake or jworsley to connect as the PostgreSQL sales user, and allows the system user named auditor to connect to PostgreSQL as either sales, or postgres.

Note: It is possible for an identifying username to be mapped to multiple PostgreSQL usernames. This is illustrated in Example 8-12 with the auditor user.

If you wish only to use ident as a means of automatically identifying your remote username, you do not need to use the pg_ident.conf file. You can instead use the special term sameuser in the pg_hba.conf file, in place of a map name.

Again, this is similar to the trusted method, however ident sameuser restricts connections based on the username provided by identd. Providing a PostgreSQL username to connect with (e.g., with the -U  flag to psql ) that is different from the name sent by identd will result in a failure to connect.

Use of the sameuser map is demonstrated in Example 8-13.

Example 8-13. A sameuser configuration

host   booktown   192.168.1.0    255.255.255.0    ident    sameuser

The host record in Example 8-13 allows any machine on the 192.168.1 network block to connect to the booktown database, using the PostgreSQL username that matches the username provided by identd. The sameuser term causes PostgreSQL to implicitly compare the requested PostgreSQL username against the name provided by identd.

Authentication Failure

When authentication failure occurs, PostgreSQL will usually do its best to provide a useful error message, rather than blindly fail. The following are common error messages you may encounter, with explanations:

FATAL 1: user "testuser" does not exist

The specified username was not found in the pg_shadow system table, meaning the user does not exist. See Chapter 10 for more on adding users.

FATAL 1: Database "testdb" does not exist in the system catalog

This database cannot be found because it does not exist. Note that if you do not specify a database name to a PostgreSQL connection, it will attempt to connect to the provided username.

No pg_hba.conf entry for host 123.123.123.1, user testuser, database testdb

You have succeeded in contacting the server, but the server is not accepting your connection. The server refused the connection because it cannot find an entry for testuser using testdb at their IP address (123.123.123.1) in the pg_hba.conf file.

Password authentication failed for user 'testuser'

You have succeeded in contacting the server and it is replying back, but the connection failed password authorization. Check the password you are supplying to the server, and make sure that it is correct. Further, you can check the Kerberos or Ident software programs if you are using them for your password authentication.

You may want to check if this user has a password. If this user does not have one, and the pg_hba.conf file is set to check for passwords, it will still check every user for their password. For all users without a defined password, a NULL password is assigned to that user. When the user tries to log in and does not specify a password, it will compare the NULL password to the NULL input, and it will return false.

On the other hand, if the user tries to supply a password (even a blank one), it will compare that input value with the NULL password and still return false. If you are using password authentication, you must assign a password to all users. If a password is not assigned to a user in such a scheme, password authentication will always fail, and the user will not be able to log in.