PPoossttffiixx BBeerrkkeelleeyy DDBB HHoowwttoo

-------------------------------------------------------------------------------

IInnttrroodduuccttiioonn

Postfix uses databases of various kinds to store and look up information.
Postfix databases are specified as "type:name". Berkeley DB implements the
Postfix database type "hash" and "btree". The name of a Postfix Berkeley DB
database is the name of the database file without the ".db" suffix. Berkeley DB
databases are maintained with the postmap(1) command.

Note: Berkeley DB version 4 is not supported by Postfix versions before 2.0.

This document describes:

 1. How to build Postfix without Berkeley DB support even if the system comes
    with Berkeley DB.

 2. How to build Postfix on systems that normally have no Berkeley DB library.

 3. How to build Postfix on BSD or Linux systems with multiple Berkeley DB
    versions.

 4. How to tweak performance.

 5. Missing pthread library trouble.

BBuuiillddiinngg PPoossttffiixx wwiitthhoouutt BBeerrkkeelleeyy DDBB ssuuppppoorrtt eevveenn iiff tthhee ssyysstteemm ccoommeess wwiitthh
BBeerrkkeelleeyy DDBB

Note: The following instructions apply to Postfix 2.9 and later.

Postfix will normally enable Berkeley DB support if the system is known to have
it. To build Postfix without Berkeley DB support, build the makefiles as
follows:

    % make makefiles CCARGS="-DNO_DB"
    % make

This will disable support for "hash" and "btree" files.

BBuuiillddiinngg PPoossttffiixx oonn ssyysstteemmss tthhaatt nnoorrmmaallllyy hhaavvee nnoo BBeerrkkeelleeyy DDBB lliibbrraarryy

Some UNIXes ship without Berkeley DB support; for historical reasons these use
DBM files instead. A problem with DBM files is that they can store only limited
amounts of data. To build Postfix with Berkeley DB support you need to download
and install the source code from http://www.oracle.com/database/berkeley-db/.

Warning: some Linux system libraries use Berkeley DB, as do some third-party
libraries such as SASL. If you compile Postfix with a different Berkeley DB
implementation, then every Postfix program will dump core because either the
system library, the SASL library, or Postfix itself ends up using the wrong
version.

The more recent Berkeley DB versions have a compile-time switch, "--with-
uniquename", which renames the symbols so that multiple versions of Berkeley DB
can co-exist in the same application. Although wasteful, this may be the only
way to keep things from falling apart.

To build Postfix after you installed the Berkeley DB from source code, use
something like:

    % make makefiles CCARGS="-DHAS_DB -I/usr/local/BerkeleyDB/include" \
        AUXLIBS="-L/usr/local/BerkeleyDB/lib -ldb"
    % make

Solaris needs this:

    % make makefiles CCARGS="-DHAS_DB -I/usr/local/BerkeleyDB/include" \
        AUXLIBS="-R/usr/local/BerkeleyDB/lib -L/usr/local/BerkeleyDB/lib -ldb"
    % make

The exact pathnames depend on the Berkeley DB version, and on how it was
installed.

Warning: the file format produced by Berkeley DB version 1 is not compatible
with that of versions 2 and 3 (versions 2 and 3 have the same format). If you
switch between DB versions, then you may have to rebuild all your Postfix DB
files.

Warning: if you use Berkeley DB version 2 or later, do not enable DB 1.85
compatibility mode. Doing so would break fcntl file locking.

Warning: if you use Perl to manipulate Postfix's Berkeley DB files, then you
need to use the same Berkeley DB version in Perl as in Postfix.

BBuuiillddiinngg PPoossttffiixx oonn BBSSDD ssyysstteemmss wwiitthh mmuullttiippllee BBeerrkkeelleeyy DDBB vveerrssiioonnss

Some BSD systems ship with multiple Berkeley DB implementations. Normally,
Postfix builds with the default DB version that ships with the system.

To build Postfix on BSD systems with a non-default DB version, use a variant of
the following commands:

    % make makefiles CCARGS=-I/usr/include/db3 AUXLIBS=-ldb3
    % make

Warning: the file format produced by Berkeley DB version 1 is not compatible
with that of versions 2 and 3 (versions 2 and 3 have the same format). If you
switch between DB versions, then you may have to rebuild all your Postfix DB
files.

Warning: if you use Berkeley DB version 2 or later, do not enable DB 1.85
compatibility mode. Doing so would break fcntl file locking.

Warning: if you use Perl to manipulate Postfix's Berkeley DB files, then you
need to use the same Berkeley DB version in Perl as in Postfix.

BBuuiillddiinngg PPoossttffiixx oonn LLiinnuuxx ssyysstteemmss wwiitthh mmuullttiippllee BBeerrkkeelleeyy DDBB vveerrssiioonnss

Some Linux systems ship with multiple Berkeley DB implementations. Normally,
Postfix builds with the default DB version that ships with the system.

Warning: some Linux system libraries use Berkeley DB. If you compile Postfix
with a non-default Berkeley DB implementation, then every Postfix program will
dump core because either the system library or Postfix itself ends up using the
wrong version.

On Linux, you need to edit the makedefs script in order to specify a non-
default DB library. The reason is that the location of the default db.h include
file changes randomly between vendors and between versions, so that Postfix has
to choose the file for you.

Warning: the file format produced by Berkeley DB version 1 is not compatible
with that of versions 2 and 3 (versions 2 and 3 have the same format). If you
switch between DB versions, then you may have to rebuild all your Postfix DB
files.

Warning: if you use Berkeley DB version 2 or later, do not enable DB 1.85
compatibility mode. Doing so would break fcntl file locking.

Warning: if you use Perl to manipulate Postfix's Berkeley DB files, then you
need to use the same Berkeley DB version in Perl as in Postfix.

TTwweeaakkiinngg ppeerrffoorrmmaannccee

Postfix provides two configuration parameters that control how much buffering
memory Berkeley DB will use.

  * berkeley_db_create_buffer_size (default: 16 MBytes per table). This setting
    is used by the commands that maintain Berkeley DB files: postalias(1) and
    postmap(1). For "hash" files, create performance degrades rapidly unless
    the memory pool is O(file size). For "btree" files, create performance is
    good with sorted input even for small memory pools, but with random input
    degrades rapidly unless the memory pool is O(file size).

  * berkeley_db_read_buffer_size (default: 128 kBytes per table). This setting
    is used by all other Postfix programs. The buffer size is adequate for
    reading. If the cache is smaller than the table, random read performance is
    hardly cache size dependent, except with btree tables, where the cache size
    must be large enough to contain the entire path from the root node.
    Empirical evidence shows that 64 kBytes may be sufficient. We double the
    size to play safe, and to anticipate changes in implementation and bloat.

MMiissssiinngg pptthhrreeaadd lliibbrraarryy ttrroouubbllee

When building Postfix fails with:

    undefined reference to `pthread_condattr_setpshared'
    undefined reference to `pthread_mutexattr_destroy'
    undefined reference to `pthread_mutexattr_init'
    undefined reference to `pthread_mutex_trylock'

Add the "-lpthread" library to the "make makefiles" command.

    % make makefiles .... AUXLIBS="... -lpthread"

More information is available at http://www.oracle.com/database/berkeley-db/.