DESCRIPTION:
------------

This is version 1.09 of Apache::AuthDBI and Apache::DBI.

These modules are supposed to be used with the Apache server together with 
an embedded perl interpreter like mod_perl. They provide support for basic 
authentication and authorization as well as support for persistent database 
connections via Perl's Database Independent Interface (DBI). 

o DBI.pm provides persistent database connections: 
  - connections can be established during server-startup 
  - configurable rollback to ensure data integrity 
  - configurable verification of the connections to avoid time-outs. 

o AuthDBI.pm provides authentication and authorization: 
  - optional shared cache for passwords to minimize database load 
  - configurable cleanup-handler deletes outdated entries from the cache 

Apache::DBI has been in widespread deployment on many platforms for
years.  Apache::DBI is one of the most widely used mod_perl related
modules.  It can be considered stable.



RECENT CHANGES:
---------------

  See the Changes file for more detail


DEVELOPMENT:
------------

Apache::DBI is in svn at perl.org; see
http://svn.perl.org/modules/Apache-DBI


EXAMPLES:
---------

Here we explain only some simple examples. For further information and 
limitations please read the module documentation. 


1. user authentication

Suppose you want to restrict access to a certain URL to a specific user and 
the necessary information for restricting user access is stored in your 
database. A typical setup would be the following: 

conf/httpd.conf:

  PerlModule Apache::AuthDBI

URL/.htaccess:

  AuthName DBI
  AuthType Basic

  PerlAuthenHandler Apache::AuthDBI::authen

  PerlSetVar Auth_DBI_data_source   dbi:driver:dsn
  PerlSetVar Auth_DBI_username      db_username
  PerlSetVar Auth_DBI_password      db_password
  # DBI->connect($data_source, $username, $password)

  PerlSetVar Auth_DBI_pwd_table     users
  PerlSetVar Auth_DBI_uid_field     username
  PerlSetVar Auth_DBI_pwd_field     password
  #SELECT pwd_field FROM pwd_table WHERE uid_field=$user

  require user myuser

In this example it is assumed, that your database contains a table named 
'users' which has at least the two columns 'username' and 'password'. When 
accessing the URL for the first time a requester pops up, asking for username 
and password. For authentication the module retrieves for the given username 
the password from the database. This is compared with the crypted password 
given by the user. If the check succeeds, the user is given access to the 
specified URL. 

Please do not confuse this user authentication with the username/password 
needed for the database connect. These two authentications are completely 
independent !

Windows users should turn off the case-sensitive option.


2. group authorization

Suppose you want to restrict access to a certain URL to a specific user group 
and the necessary information for restricting user access is stored in your 
database. A typical setup would be the following: 

conf/httpd.conf:

  PerlModule Apache::AuthDBI

URL/.htaccess:

  AuthName DBI
  AuthType Basic

  PerlAuthenHandler Apache::AuthDBI::authen
  PerlAuthzHandler  Apache::AuthDBI::authz

  PerlSetVar Auth_DBI_data_source   dbi:mydriver:mydsn
  PerlSetVar Auth_DBI_username      db_username
  PerlSetVar Auth_DBI_password      db_password
  # DBI->connect($data_source, $username, $password)

  PerlSetVar Auth_DBI_pwd_table     users
  PerlSetVar Auth_DBI_uid_field     username
  PerlSetVar Auth_DBI_pwd_field     password
  PerlSetVar Auth_DBI_grp_field     groupname
  #SELECT grp_field FROM pwd_table WHERE uid_field=$user

  require group mygroup

In this example it is assumed, that your database contains a table named 
'users' which has at least the three columns 'username', 'password' and 
'groupname'. When accessing the URL for the first time a requester pops up, 
asking for username and password. The first check (authentication) retrieves 
for the given username the password from the database. This is compared with 
the crypted password given by the user. In a second check (authorization) 
the groups of the given username are looked up in the database and compared 
with the groups required in the .htaccess file. If both checks succeed, the 
user is given access to the specified URL. 

Please do not confuse the user authentication with the username/password 
needed for the database connect. These two authentications are completely 
independent ! 

Although authorization handles all types of basic authentication it is 
perfectly sufficient to configure only authentication, as long, as the 
require token restricts access to 'valid-user' or to one or more single user 
names. You need to configure authorization only if you have more than one 
require token or if the require token contains one or more group names. 


3. persistent database connection

The following information is intended to motivate the use of persistent 
database connections and to explain the necessary configuration. 

In the above example for user authorization the requester asking for username 
and password pops up only once. The browser stores the user input and provides 
it to subsequent requests. But the sequence of two database accesses is done 
for every request, e.g. if your restricted URL contains a HTML page with some 
images, this sequence is executed once for the HTML page and once for every 
image ! For databases which needs a significant amount of time for the connect 
(e.g. start of a backend process) this might become an unacceptable overhead 
for the authorization procedure. This drawback can be overcome with the use of 
persistent database connections as provided by the Apache::DBI module. 

The benefit of a persistent database connection is not limited to the use 
of authorization. Every application, which does a lot of database queries, 
should gain a significant performance boost, when using persistent database 
connections. 

If you plan to use persistent database connections, there is only one thing 
to do: add the following configuration directive to conf/httpd.conf or to 
your startup.pl:

  PerlModule Apache::DBI    # this comes first !!
  ....                      # other modules using DBI

Do not change your perl scripts ! In particular do not add any 
'use Apache::DBI;' statements. Also there is no need to remove 
the $dbh->disconnect statements from your perl scripts. 

The DBI module checks when it is loaded if the Apache::DBI module has been 
loaded before (that's the reason the Apache::DBI module has to come first). 
In this case, during the database connect, control flow goes through the 
Apache::DBI module which stores the new database handle in a global hash and 
which overloads the disconnect method with a do-nothing. 

With the above configuration every server initiates a database connection upon 
the first connect request. Sometimes it is more convenient to initiate all 
needed database handles upon process startup. This can be done with the method: 

 Apache::DBI->connect_on_init($data_source, $username, $auth, \%attr)

This method is supposed to be called in a startup file, in which also all 
needed modules can be loaded. As an example the file startup.pl is provided. 
Add all other modules you need to this file and just add one line to your 
httpd.conf:

 PerlRequire /usr/local/apache/perl/startup.pl

This way all modules are pulled into the main httpd process. When the main 
process forks his children, the code of all modules is already in place and 
the database handle will also be initiated. 

WARNING: Do not attempt to open a persistent database connection in the parent 
process (via PerlRequire or PerlModule). If you do, children will get a copy 
of this handle, causing clashes when the handle is used by two processes at 
the same time. Each child must have it's own unique connection handle. For 
the same reason it is not possible, to share one database handle between all 
servers using some IPC mechanism. 

If you want to make sure that the module works correctly, turn on debugging 
as described below and search for 'Apache::DBI' in the output. You should 
get one 'new connect' message for every server process. Any subsequent request 
should result in a 'already connected' message. Please keep in mind, that 
server processes may be killed as well as newly created depending upon your 
configuration and depending upon your load. Every new server process needs to 
do its own initial database connect. 

Another useful method for enhancing the performance is to enable the caching in 
AuthDBI setting Auth_DBI_cache_time > 0 and to use shared memory for the cache 
(see the module documentation for details). This will reduce the database load 
considerably. 


COPYRIGHT:
----------

You may distribute under the terms of either the GNU General Public
License or the Artistic License, as specified in the Perl README file.


PREREQUISITES:
--------------

Configure mod_perl1 with: 

  perl Makefile.PL PERL_CHILD_INIT=1 PERL_AUTHEN=1 PERL_AUTHZ=1 
                   PERL_CLEANUP=1 PERL_STACKED_HANDLERS=1

If there are no security reasons to limit the API, just use EVERYTHING=1. 

mod_perl2 RC5 and higher should work with Apache::DBI 0.96 and higher.
No specific switches must be passed to mod_perl2's Makefile.PL.

INSTALLATION:
-------------

   perl Makefile.PL
   make
   make test      # only works with MySQL so far; patches welcome
   make install



IF YOU HAVE PROBLEMS:
---------------------

Please read the README and the the module documentation: 'perldoc Apache::AuthDBI', 
'perldoc Apache::DBI'.
Please verify your setup: turn on debug output and compare it to traces.txt. 

If you have problems with persistent database connections, verify that everything 
works correct without using Apache::DBI.

Before sending a bug report it might be useful to look at the debug output. 
To enable full debug output set the following variables in startup.pl or in your 
perl script: 

  $Apache::DBI::DEBUG     = 2;
  $Apache::AuthDBI::DEBUG = 2;

and watch the error_log. Compare the output to the traces in traces.txt.

If this doesn't help, please send an email to <modperl@apache.org> and include 
the following information in your bug-report:

 - debug output,
 - output of perl -V,
 - version of ApacheDBI,
 - version of DBI,
 - used database


A common problem is an error-message that $dbh will not stay shared. A 
complete explanation for this behavior is given in the modperl-FAQ. In 
short, instead of this:

 my $dbh = ...;
 subroutine();
 sub subroutine {
   $dbh->....
 }

do this:

 my $dbh = ...;
 subroutine($dbh);
 sub subroutine {
   my $dbh = shift;
   $dbh->....
 }



FURTHER INFORMATION:
--------------------

mod_perl by Doug MacEachern modperl-subscribe@perl.apache.org
                            http://perl.apache.org/

DBI      by Tim Bunce       dbi-users-subscribe@perl.org
                            http://dbi.perl.org/

Apache   by Apache Group    news:comp.infosystems.www.servers.unix
                            users-subscribe@httpd.apache.org
                            http://httpd.apache.org/


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

   Edmund Mergl <E.Mergl@bawue.de>
   Ask Bjoern Hansen <ask@develooper.com>
   Philip M. Gollucci <pgollucci@p6m7g8.com>

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