OpenLDAP

From Gentoo Wiki
Jump to:navigation Jump to:search
Resources

OpenLDAP is a free LDAP server implementation offering an X.500 directory service. The directory service enables user or service lookup, permission management, centralized login, key infrastructure and much more. It includes a suite of LDAP utilities and a standalone LDAP server daeomon.

Installation

USE flags

USE flags for net-nds/openldap LDAP suite of application and development tools

+berkdb Add support for sys-libs/db (Berkeley DB for MySQL)
+cleartext Enable use of cleartext passwords
+syslog Enable support for syslog
argon2 Enable password hashing algorithm from app-crypt/argon2
autoca Automatic Certificate Authority overlay
crypt Add support for encryption -- using mcrypt or gpg where applicable
cxx Build support for C++ (bindings, extra libraries, code generation, ...)
debug Enable extra debug codepaths, like asserts and extra output. If you want to get meaningful backtraces see https://wiki.gentoo.org/wiki/Project:Quality_Assurance/Backtraces
experimental Enable experimental backend options
gnutls Prefer net-libs/gnutls as SSL/TLS provider (ineffective with USE=-ssl)
iodbc Add support for iODBC library
ipv6 Add support for IP version 6
kerberos Add kerberos support
kinit Enable support for kerberos init
minimal Build libraries & userspace tools only. Does not install any server code
odbc Enable ODBC and SQL backend options
overlays Enable contributed OpenLDAP overlays
pbkdf2 Enable support for pbkdf2 passwords
perl Add optional support/bindings for the Perl language
samba Add support for SAMBA (Windows File and Printer sharing)
sasl Add support for the Simple Authentication and Security Layer
selinux !!internal use only!! Security Enhanced Linux support, this must be set by the selinux profile or breakage will occur
sha2 Enable support for pw-sha2 password hashes
smbkrb5passwd Enable overlay for syncing ldap, unix and lanman passwords
ssl Add support for SSL/TLS connections (Secure Socket Layer / Transport Layer Security)
static-libs Build static versions of dynamic libraries as well
systemd Enable use of systemd-specific libraries and features like socket activation or session tracking
tcpd Add support for TCP wrappers
test Enable dependencies and/or preparations necessary to run tests (usually controlled by FEATURES=test but can be toggled independently)
Data provided by the Gentoo Package Database · Last update: 2024-09-12 18:37 More information about USE flags

Emerge

Recommended USE flags:

  • any of argon2, pbkdf2, or sha2 to provide modern algorithms for secure hashing of passwords.
  • sasl to bootstrap configuration through Simple Authentication and Security Layer.
  • ssl to encrypt traffic over the network so that passwords stay safe.
  • syslog to log events, e.g. failed login attempts.
  • debug to enable use of -d flag.


Tip
The utilities installed by OpenLDAP do not always provide helpful error messages. It is recommended that the debug USE flag is enabled during the initial setup. Later, it can be disabled and OpenLDAP rebuilt without it, if desired.
root #emerge --ask net-nds/openldap


In addition to the Standalone LDAP Daemon (slapd), OpenLDAP installs two sets of utilities. Those prefixed with ldap* (ldapadd, ldapmodify, ldapsearch, ...) and those prefixed with slap* (slapadd, slapmodify, slaptest, ...).

  • The ldap* commands open a connection to a running ldap server (slapd or otherwise) and perform LDAP operations. They will not work if the LDAP server is not running.
  • The slap* commands operate on an OpenLDAP database stored somewhere on the filesystem. They should not be used to modify a database that is in use by a running instance of slapd. Even though the slap* commands may sometimes appear to succeed if they are run on a live database, they can actually interfere with normal slapd operation and potentially cause data corruption!

Configuration

Prior to OpenLDAP 2.3, slapd and all of the slap* utilities were configured from a single file /etc/openldap/slapd.conf. Since version 2.3, a new method of configuration was added and the old file was deprecated. The configuration is now stored inside an LDAP directory which (by default) is stored at /etc/openldap/slapd.d/. On a fresh install this directory will not yet exist and must be created before slapd can be used.

The ldap* utilities were and still are configured from a single file /etc/openldap/ldap.conf. This file may need to be edited before the ldap* utilities can connect to an LDAP server secured with TLS. See the note in the TLS section.

Note
As of OpenLDAP version 2.6, the slapd.conf file can still be used, but since it is deprecated, its use will not be covered by this wiki page. See man 5 slapd.conf for details.

LDIF Files

The configuration directory must be initialized from an LDAP Data Iinterchange Format (LDIF) file. A few files are provided by the package: /etc/openldap/slapd.ldif and /etc/openldap/slapd.default.ldif. These files will be used as a starting point to configure the daemon.

Because the configuration of slapd is itself an LDAP directory, the configuration data is arranged into a Directory Information Tree (DIT) just like a "real" LDAP directory. The "path" of an entry in the tree is called its "Distinguished Name" abbreviated dn. In a DIT, the paths can branch out similar to the paths of a filesystem except the root of the tree is on the right instead of on the left. For Example, the hypothetical Distinguished Name uid=Larry,ou=Users,dc=gentoo,dc=org would correspond to hypothetical filesystem path /org/gentoo/Users/Larry. In an LDIF file, entries are defined by a line stating the dn followed by lines describing its attributes and their values. Here is an example entry for Larry the cow:

FILE larry.ldif
# Entry for Larry the Cow!
dn: uid=larry25,ou=Users,dc=gentoo,dc=org
uid: larry25
cn: Larry the Cow
givenName: Larry
sn: Cow
objectClass: inetOrgPerson
objectClass: organizationalPerson
objectClass: person

Each entry in the DIT can contain any number of attribute/value pairs. It can also have the same attribute listed multiple times with different values if the schema allows it. This is known as a multi-valued attribute. The special attribute objectClass defines the schema the entry must follow. It is possible for an entry to have more than one objectClass but only if they satisfy certain conditions. See section Schema.

Note
A single LDIF file can contain several entries. Entries must be separated by at least one blank line. If a blank line were to be inserted in the middle of an entry to improve readability, it may cause a parsing error. Comment lines (lines starting with #) don't count, so they can be freely inserted in the middle of an entry, but care should be taken to ensure that two separate entries are not accidentally connected by comment lines.

For more information about LDIF files, see man 5 ldif.

Creating slapd.d

In the provided slapd.ldif, the config tree is defined with default entries containing default attributes and values. The different entries are as follows:

  • dn: cn=config This entry contains global configuration of the daemon
  • dn: cn=module,cn=config This entry contains the configuration of the daemon's loadable modules
  • dn: cn=schema,cn=config This entry is followed by include directives which define the schemas loaded by the daemon
  • dn: olcDatabase=frontend,cn=config This entry contains database settings which are applied to all databases served by the daemon
  • dn: olcDatabase=config,cn=config This entry represents the config database itself. Settings can be changed here that allow access to the config database so that the daemon's configuration can be changed while it is running.
  • dn: olcDatabase=mdb,cn=config This entry contains the settings for the primary database served by the daemon. It is possible to define additional entries like this so that the daemon hosts more than one database, but that is beyond the scope of this document.


The configuration of the entries is determined by the attributes that begin with "olc" e.g. olcAccess: The different configuration options and what they mean are described in the slapd-config(5) manpage.

Firstly, the olcPidFile and olcArgsFile attributes need to be changed. This is because the daemon is not run as root on Gentoo (by default). The daemon needs to access the pid and args files after it drops privileges and for that it needs a directory under /run. Start by opening slapd.ldif in a text editor.

FILE /etc/openldap/slapd.ldif(Excerpt)
#
# Change these from their defaults!
olcArgsFile: /run/openldap/slapd.args
olcPidFile: /run/openldap/slapd.pid

Next, the dn: olcDatabase=mdb,cn=config entry needs to be edited. Change the olcSuffix to match the domain that the LDAP server will be accessible at. In theory, the suffix does not need to be a domain name. It is possible to have an organization as a suffix e.g. olcSuffix: o=Larry Cow Corp, but that will not be covered here.

FILE /etc/openldap/slapd.ldif(Excerpt)
dn: olcDatabase=mdb,cn=config
objectClass: olcDatabaseConfig
objectClass: olcMdbConfig
olcDatabase: mdb
#
# The max size of the DB is not super important now, but it must be defined
olcDbMaxSize: 1073741824
#
# Edit the Domain Components (dc) to match your domain. e.g. www.example.com -> dc=www,dc=example,dc=com
olcSuffix: dc=my-domain,dc=com
#
# The 'RootDN' is like the root account of your database.
# Change the Common Name (cn) here to change the name of that account (optional)
# Change the Domain Components (dc) to match your domain.
olcRootDN: cn=Manager,dc=my-domain,dc=com
#
# The 'RootPW' is the password for the RootDN.
# By default, the word "secret" is literally the password.
# This can be changed to a hashed password later on.
olcRootPW: secret
#
# Set this to where you would like to store your database files
# The database directory MUST exist prior to running slapd AND 
# should only be accessible by the slapd and slap tools.
# Mode 700 recommended.
olcDbDirectory: /var/openldap-data

As the comments in the ldif file state, the database directory must be created before proceeding. Changing the mode is also recommended

root #mkdir /var/openldap-data
root #chown ldap:ldap $_
root #chmod 0700 $_

This is the minimum config needed to run the daemon. However, if you would like to take advantage of the feature that allows you to change the daemon configuration at runtime, then the config database must be defined in the LDIF file and made accessible by adding a RootDN and password to it. The order that the entries appear in this ldif file is important. The following excerpt, if it is to be used, must be inserted before the primary database entry described in the previous excerpt. Normally order does not matter in LDAP data, but this data is special because it is the OpenLDAP config.

FILE /etc/openldap/slapd.ldif(Excerpt)
#
# This entry does not exist in the default slapd.ldif.
# You must create it if you want to access and modify the config during runtime.
# This is optional, but without it the daemon will have to be shutdown to change config
dn: olcDatabase=config,cn=config
objectClass: olcDatabaseConfig
olcDatabase: config
#
# Define the root account of the config database
olcRootDN: cn=Manager,cn=config
olcRootPW: secret
#
# The olcAccess rules do not apply to the rootDN.
# This line means that ONLY the rootDN will be allowed to connect to the database.
olcAccess: to * by * none

Now it is time to initialize the config database on the filesystem. The slapadd tool will be used for this purpose. The flag -F must point to the desired location of the slapd.d directory. Since it is possible for a slapd instance to host more than one database at once, the database that will be edited must be specified using the -n 0 flag. The databases are given indices starting at -1 and the config database being created in this step is always index 0. The flags -v verbose and -u dry-run are recommended to ensure there are no errors in the LDIF files before the tool actually writes to the directory. If a partial write occurs because of an error, then all the files in the slapd.d directory will have to be removed before trying again.

root #slapadd -v -n 0 -F /etc/openldap/slapd.d -l /etc/openldap/slapd.ldif

The files created by slappadd will have the ownership of the user that invoked it (usually root). This MUST be changed before starting the daemon.

root #chown -R ldap:ldap /etc/openldap/slapd.d
Tip
The OpenLDAP utilities are very strict about formatting and schema. If you are receiving a vague error message when trying to parse an LDIF file, try adding the debug flag -d 1. This flag is not usable unless OpenLDAP was merged with the debug USE flag.

Running the daemon

OpenRC

Before starting the daemon, edit the file /etc/conf.d/slapd. Comment and uncomment the lines defining OPTS_CONF so that the daemon is passed the capital -F option with the path to the directory instead of the lowercase -f.

FILE /etc/conf.d/slapd(Excerpt)
# Comment this line to disable the old slapd.conf file
#OPTS_CONF="-f /etc/${INSTANCE}/slapd.conf"
# Uncomment this to use the new slapd.d configuration directory
OPTS_CONF="-F /etc/${INSTANCE}/slapd.d"

The daemon can now be started.

root #rc-service slapd start

If it launches successfully, it can be added to the default runlevel so that it will start with the system if desired.

root #rc-update slapd default

If the daemon does not start because of an error, it can be run manually with the debug flags enabled to diagnose the problem.

root #/usr/lib64/openldap/slapd -d 1 -h ldap:// -F /etc/openldap/slapd.d/ -u ldap -g ldap

Systemd

TODO

Management

Creating the first entry

The top-most entry in an LDAP database is called the Root DSE (Not to be confused with the Root DN). This entry is special in the LDAP protocol and its attributes provide information about the server to any clients that connect. The root DSE is not meant to hold any directory data, just metadata. It cannot be changed at the protocol level and should be ignored for most purposes. Immediately below the root DSE is the DIT (or DITs, there can be more than one) which is where the first "real" entry lives. All of the entries in the tree are descendants of this first entry. slapd does not create this entry by default, and most LDAP tools will be confused if it doesn't exist so it should be created now.

First, a new ldif file must be created with a single entry. The dn of the entry must match the olcSuffix which was set earlier in /etc/openldap/slapd.ldif. The attributes of the entry can be set to almost anything according to the needs of the database administrator, but the entry must have a STRUCTURAL objectClass defined, and it must satisfy the schema of that object class. Schemas and the different types of object classes are discussed later in the schema section. Here is an example first entry that will work:

FILE first.ldifExample First Entry
# Change the Domain Components (dc) to match your domain from earlier
dn: dc=example,dc=com
objectclass: dcObject
objectclass: organization
#
# In this example, Organization (o) is used as the STRUCTURAL objectClass.
# The value of the o attribute can be anything
o: Larry the Cow Example Organization
#
# This dc must match the first dc in the list above.
# The rest of the dc's in the list can be omitted
dc: example

With first.ldif saved to the working directory and slapd already running, the utility ldapadd can be used to create the first entry. It does not need to be run as root. After the -D option, the RootDN must be specified. This is the root account name from slapd.ldif when the database was created. Running the command will prompt for the password which was also set in slapd.ldif.

user $ldapadd -v -D "cn=Manager,dc=example,dc=com" -W -H ldap://localhost -f dse.ldif

Modifying the tree

At this point, the slapd server is running and configured well enough that further changes to the database can be made using a tool like Apache Directory Studio, net-nds/shelldap, or net-nds/phpldapadmin. Using tools like these is highly recommended, because editing ldif files by hand and parsing them with command line tools is tedious and error-prone. Nevertheless, it may become necessary to use the command line tools from OpenLDAP for one reason or another, so this section will discuss using them.

In general there are two methods of editing an OpenLDAP database:

  • Using the ldap* tools to upload an ldif file to a running instance of slapd.
  • Shutting down slapd and using the slap* tools to apply an ldif file directly on the database files.

The latter method has some disadvantages. If the database resides in a secure location (as it should), then the change commands must be run as root. After changes are made, it is important to ensure the permissions of the database files are correct. When slapadd and slapmodify are run as root, they may change the ownership of the database files to root. After the permissions are corrected, slaptest can be used to check the database.

Both methods require the creation of LDIF-formatted text. It can be generated and piped to the commands, or saved to a file that is passed as a command argument.

Change records

Adding a new entry is simple and requires exactly the same process as adding the first entry. Editing an existing entry requires creating an LDIF Change Record. Change records are almost exactly the same as the LDIF records used previously, but they have a few additional special attributes.

FILE change.ldifExample Change Record
# DN of the entry to be changed
dn: uid=larry,dc=example,dc=com
#
# What type of change is happening: (modify)
changeType: modify
#
# Which attribute(s) are being modified: (givenName)
replace: givenName
#
# New definitiion(s) of the specified attribute
givenName: Lawrence

The valid changeType's and their behavior are described further in the LDIF(5) manpage. The change can be applied with ldapmodify or slapmodify.

user $ldapmodify -v -D "cn=Manager,dc=example,dc=com" -W -H ldap://localhost -f change.ldif

OR

root #<service manager> stop slapd
root #slapmodify -v -n 1 -l change.ldif
root #chown -R ldap:ldap /var/openldap-data
root #slaptest
root #<service manager> start slapd

Managing Passwords

In LDAP, the act of authentication is known as binding. Almost any entry in the tree can be bound to, not just users. Typically, all an entry needs to be bound is an attribute called userPassword which stores the password used during the bind (The schema must also allow the entry to have this attribute). If the cleartext USE flag was enabled, then OpenLDAP will allow the userPassword to store the password literally, but this is highly insecure and should only be used for testing. Instead, a password hashing algorithm should be used.

OpenLDAP supports several password hashing algorithms out-of-the-box, and more can be enabled through modules which are installed by USE flags.

Warning
Even if a strong password hashing & salting scheme is used, it will still not offer perfect protection against any attempts at password cracking. Care must be taken to prevent anyone or anything from reading the LDAP database files. These files should have the most restrictive permissions possible while still allowing the daemon to function.

An entry with a hashed password looks like this:

FILE hashed_password.ldifExample Hashed Password
dn: uid=lcow,ou=users,dc=example,dc=com
objectClass: person
objectClass: posixAccount
objectClass: top
homeDirectory: /home/lcow
cn: Larry the Cow
sn: Cow
uid: lcow
uidNumber: 1001
gidNumber: 1000
# Typical textual representation of a hashed password
userPassword: {SSHA}LryXSzUzmhjvY99K9wWC1iPE6TyRTeme

# Sometimes the entry will be displayed as hexadecimal.
# In this case the attribute will have two colons
# userPassword:: 61615d61a81556ab82c8f4ea0573eb03c048cf33

The value inside the braces, SSHA in this case, is the hashing/salting algorithm. The text immediately following is the hashed data. In order to generate the userPassword value, the slappasswd tool can be used. -h specifies the desired hashing algorithm.

user $slappasswd -h '{SSHA}'
New password:
Re-enter new password:
{SSHA}McNjDs0vQlBpfR8NKFttwTBBb0awnOJR

This output can be saved to a file, or piped to another command to change a password. Here is an example script that prompts for a password, and then generates a new LDIF record which is piped to slapmodify to change the password of the Root DN.

CODE Example script for changing the RootDN password of an offline database
#!/bin/bash

# This example uses the ARGON2 algorithm which must be loaded from a module
# If a different builtin algorithm is used, the "-o module-path=..." and "-o module-load=..." options can be removed
HASH_ALGO='{ARGON2}'
HASH_MODULE='argon2.so'

{
        # This dn is from the config database
        echo   'dn: olcDatabase={1}mdb,cn=config'
        
        # Generate an LDIF change record
        echo   'changeType: modify'
        echo   'replace: olcRootPW'
        printf "olcRootPW: %s\n" \
              $(slappasswd -o module-path=/usr/lib64/openldap/openldap -o module-load="${HASH_MODULE}" -h "${HASH_ALGO}")

# Pipe the change record with the new hashed password into slapmodify
# Use "-n 0" because this is a modification to the config database
} | slapmodify -v -n 0

In the previous example, a hash algorithm is used which is not built into OpenLDAP. It must first be enabled by a USE flag and then loaded as a module. The module can be loaded by slappasswd on the command line, but in order for slapd to understand the format of the entry, it must have the module added to its configuration database. See TODO (Advanced Config/Modules).

Advanced configuration

TLS

Transport Layer Security (TLS) can be enabled on slapd to ensure that communication between clients and the daemon is encrypted and safe from eavesdropping. This is absolutely necessary if the LDAP server is going to store any personal or security-sensitive information, including passwords.

slapd supports implicit (ldaps://) and explicit (STARTTLS) TLS.

  • Implicit TLS can only be enabled by passing the -h ldaps:// command line argument to slapd when it is started. The Gentoo init scripts do this by default, but this can be changed by editing /etc/conf.d/slapd. When implicit TLS is enabled, slapd will listen for connections on port 636 by default.
  • Explicit TLS is always an option by default. Its use must be initiated by the client. It works when clients connect to the normally unencrypted LDAP port 389 and issue the STARTTLS command to upgrade the connection to an encrypted channel. Explicit TLS will not be used if unencrypted connections are completely disabled in /etc/conf.d/slapd. For more information on this behavior, see man 8 slapd.

Before TLS will work, slapd must be configured with a certificate and private key. For information on creating and managing certificates, see Certificates and Certificates/Become your own CA. Once a certificate and key are stored in a safe place on the filesystem, slapd can be configured to use them by adding a few options to the configuration database. It does not seem to be possible to add these entries while the daemon is running. A few different methods of updating the config database have already been demonstrated in this article. One of the previous methods could be used, but a different method using bash scripting will be demonstrated here:

CODE Example script to add or update TLS certificates
#!/bin/bash

# Set these variables to the paths of your certificate and key
CERT_PATH="/path/to/domainCert.pem"
PRIVKEY_PATH="/path/to/private.key.pem"

# To set the attributes for the first time, use 'add'
# To update the attributes later, use 'replace'
OP='add'

make_rec() {
        # Generate an LDIF change record for the config
        echo   'dn: cn=config'
        echo   'changeType: modify'
        printf '%s: %s\n' $1 $2
        printf '%s: %s\n' $2 $3
}

rc-service slapd stop && \
make_rec $OP 'olcTLSCertificateFile' $CERT_PATH | slapmodify -v -n 0 && \
make_rec $OP 'olcTLSCertificateKeyFile' $PRIVKEY_PATH | slapmodify -v -n 0 && \
chown -R ldap:ldap /etc/openldap/slapd.d && \
rc-service slapd start

The script can be simply edited according to the comments. Systemd users should also change the rc-service slapd start/stop lines to systemctl start/stop slapd. If the daemon does not come back up after running the script, the most likely problem is permissions of the certificate/key files. Recall that slapd is running as user ldap. If the ldap user does not have permission to access the files or their parent directories, then the daemon will not be able to read the files and will refuse to start. If necessary, the permissions should be changed, but care should be taken to ensure the private key is inaccessible by any user that does not strictly require access.

Important
The ldap* utilities included with OpenLDAP do not trust any certificate authorities by default. If they are to be used to connect to an LDAP server using any kind of TLS, the connection will fail unless they are configured to recognize the CA that signed the server's certificate. If the server has a certificate signed by a public certificate authority, then the trust can be established by adding the line TLS_CACERTDIR /etc/ssl/certs to the file /etc/openldap/ldap.conf.

Modules

The modules loaded by the daemon are stored in entries of the config tree. The first module entry (cn=module{0},cn=config) was added when the config database was initialized. Additional modules can be loaded at runtime by creating new module entries in LDIF files and adding them to the tree with ldapadd. It does not appear to be possible to remove modules once loaded without first taking the daemon offline.

FILE new_mods.ldifExample Module Entry
# It may be necessary to increment the index number after the module name if module entries have already been added
#             v
dn: cn=module{1},cn=config
objectClass: olcModuleList
#
# Add as many modules here as desired
olcModuleLoad: back_relay.so
olcModuleLoad: back_null.so
olcModuleLoad: argon2.so
# 
# The default path where modules are stored.
# If there is an error, check to make sure the module file is actually there
# Some modules are not present unless their USE flag was enabled
olcModulePath: /usr/lib64/openldap/openldap
user $ldapadd -v -D "cn=Manager,cn=config" -W -H ldap://localhost -f new_mods.ldif

Schema

Throughout this document, examples of DIT/LDIF entries have been presented with arbitrary attributes like cn, sn, and homeDirectory. LDAP and X.500 directories in general do not require entries to follow any particular pattern or structure. However, having a set of rules regarding the structure of an entry makes it much easier for people and software accessing the directory to understand the data and make changes. These rule sets are called schemas. Schemas can rigorously define which entries may/should/must have which attributes. They can also define a set of legal values that an attribute can have.

When a schema is loaded by slapd, it will be enforced, and any attempt to change an entry that violates the schema will be refused with an error message. System administrators can create arbitrary schemas and use them with slapd, but OpenLDAP also ships with a set of widely recognized schemas stored in /etc/openldap/schema/. This section will focus on the included schemas.

When in use, schemas are stored in the configuration database as child entries of the cn=schema,cn=config entry. Each schema entry contains dozens of attributes that actually describe the schema. It is possible to create schema entries by hand that will work, but this is tedious and error prone. Instead, the .ldif files in the schema directory schould be used.

The schema directory includes a README file which provides a description for each of the included schemas. Additionally, each of the .schema and .ldif files in the directory contain a header with commentary about the schema file. The .schema and .ldif files with the same prefix generally contain the same schema information. For example, msuser.schema and msuser.ldif both contain the Microsoft schema. The only difference is that the .schema files are formatted for inclusion in the legacy slapd.conf configuration file, and the .ldif files are formatted for use in the newer configuration database.

Note
Some schema files are dependent on each other and must be added in a certain order. Most, if not all, of the schemas included with OpenLDAP depend on the core.ldif schema, so it should be added first.

If all that is desired is adding new schemas, then the .ldif files in the schema directory can be fed directly into ldapadd or slapadd.

user $ldapadd -v -D "cn=Manager,cn=config" -W -H ldaps://example.com < core.ldif

If schemas need to be changed or removed, then the process is a little more involved. It does not appear to be possible to remove or edit schemas while the daemon is running.

root #sytemctl stop slapd

Once the daemon is stopped, the schemas can be changed with slapmodify. Keep in mind that removing or replacing the schema of a database that already has entries in it may cause errors if the existing entries fail to conform to the new schema.

If it is desired to completely remove all schema information from the database and "start from scratch", then the slapcat tool can be used. The purpose of this tool is to read an existing OpenLDAP database on the filesystem and output a new LDIF file containing all of its entries.

In the next example, the slapcat command will read the entire contents of the config database, except it will filter out all the schema information.

root #slapcat -n 0 -H "ldap:///???(!(entryDN:dnSubtreeMatch:=cn=schema,cn=config))" > new_schemas.ldif

The output file from the previous command can then be opened and the new schema information appended:

FILE new_schemas.ldif(Add to the end of the file)
# This entry is the parent of all other schema entries
# and it needs to come before them in the file
dn: cn=schema,cn=config
objectClass: olcSchemaConfig
cn: schema

# Use the include directive to copy the contents of the LDIF
# files in the schema directory into this file
include: file:///etc/openldap/schema/core.ldif

# Make sure blank lines are kept in-between entries
include: file:///etc/openldap/schema/cosine.ldif

include: file:///etc/openldap/schema/inetorgperson.ldif

Now that the new LDIF file is prepared with the correct schemas, it can be written back to the slapd.d/ directory. This file already includes all of the content of that directory minus the old schemas which were filtered out. First, create a backup of slapd.d/ directory in case anything goes awry, then overwrite it with slapadd:

root #mv /etc/openldap/slapd.d/ /etc/openldap/slapd.d.old
root #mkdir /etc/openldap/slapd.d
root #slapadd -n 0 -F /etc/openldap/slapd.d/ -l new_schemas.ldif
root #chown -R ldap:ldap /etc/openldap/slapd.d
root #slaptest

If the slapadd command fails with an error, it may be because the schemas added have an unsatisfied dependency. If everything succeeds without errors, then the daemon can now be restarted.

root #sytemctl start slapd

Beware, if the main database was not empty when the schema was changed, the daemon may refuse to start or emit lots of errors if the old entries do not conform to the new schema.

Security

Warning
This entire page does not provide a very secure LDAP service, and this section is still under construction.
  • access control and directory information tree could enable at least guessing entries by receiving different error codes (not found vs. invalid privileges) even without proper authentication or permission
  • ACLs allow to hide attributes, mind the defaults that often allow unauthenticated world readable attributes, sometimes they are omitted in simple queries but are visible if selected explicitely
  • unique identifiers must be really unique, never re-assigned, unrelated to a real person's personal data and immutable over the entire lifespan, last names not only change but names in general are not limited to 2 words containing only US-ASCII characters[1]
  • make regular backups and check their usability

Access Control

TODO

SASL bind

TODO

Attacks on LDAP servers

TODO

Troubleshooting

  • kill -INT `cat /run/openldap/slapd.pid` – Graceful shutdown if regular service script fails or a daemonized test run does not stop[2]

See also


Useful manpages

  • slapd(8)
  • slapd.conf(5)
  • slapd-config(5)
  • LDIF(5)
  • slapmodify(8)
  • ldapmodify(1)
  • ldap.conf(5)


Packages supporting LDAP-authentication/ user management:

  1. ↑ [1]Best Practices For Unique Identifiers
  2. ↑ [2]OpenLDAP Software 2.6 Administrator's Guide
Retrieved from "https://wiki.gentoo.org/index.php?title=OpenLDAP&oldid=1289962"