Jump to: navigation, search

Obsolete:InstallAndDeploy

Obsolete Documentation

Please note that this content was marked as obsolete. We left the content here for study and archaeological purposes.

Please check our Documentation for a recent list of topics.

Installing the IPA Server

Introduction

This page provides instructions on how to download the freeIPA server software, and to get it installed and configured on your system. It also provides information on common problems and possible solutions.

Assumptions

FreeIPA relies on a script to automate many of the installation tasks. This script makes a number of assumptions.

  • System:
That you are performing the installation on a "clean" system. The script overwrites a number of files without prompting for confirmation. Configuring the FreeIPA Server lists the services whose scripts and configuration files are modified.
  • Directory Server:
That there are no existing Directory Server instances. There is currently no automatic upgrade facility from an existing Directory Server. Any migration requirements should be referred to Red Hat Global Professional Services.
  • DNS:
    • That the server's machine name is set, and that it resolves to its public IP address (not to localhost).
    • That DNS is correctly configured to resolve forward and reverse addresses. The DNS does not need to be on the same machine as the IPA Server, but it does need to be fully functional.

Required Ports

FreeIPA makes use of the following ports:

  • TCP
    • 80, 443: HTTP/HTTPS
    • 389, 636: LDAP/LDAPS
    • 88, 464: Kerberos
  • UDP
    • 88, 464: Kerberos
    • 123: NTP

You need to ensure that these ports are available; that is, that they are not assigned to another service. If these ports are not available, IPA will not function correctly.

File Systems

The default root directory for users' home directories is /home, but it is the responsibility of the system administrator to ensure that whatever value is specified for this attribute is actually available.

Red Hat Enterprise Linux and most other Linux distributions include a pam module called pam_mkhomedir that can be used to automatically create a home directory if one does not exist for the user authenticating against the system. IPA does not force the use of this module because it may try to create home directories even when the shared storage is simply not available. It is the responsibility of the system administrator to activate this module on the clients if needed.

If a suitable directory and mechanism are not available for the creation of home directories, users may not be able to log in successfully.

DNS

It is recommended that you use DNS to facilitate Service Discovery in IPA. Service Discovery refers to the way that IPA clients find (or discover) IPA servers. You can use the basic DNS configuration that is provided with IPA to configure an existing DNS to work with IPA, or use --setup-bind option to configure a new DNS. The DNS does not need to be on the same machine as the IPA server, but it does need to be correctly configured and fully functional.

Note.pngNote:

The --setup-bind option is an optional parameter that can be passed to the ipa-server-install script. This is provided for convenience only; it is not a supported aspect of IPA. The following article may help you to configure your DNS server: How to set up a DNS server

To aid in the creation and configuration of a suitable DNS setup, the IPA installation creates a sample zone file. During the installation you will see a message similar to the following:

Sample zone file for bind has been created in /tmp/sample.zone.F_uMf4.db

You should use this file in your zone file in DNS. Further, you need to ensure that your FQDN does not resolve to your loopback address.

IPA, DNS, and NSCD

It is recommended that you avoid or restrict the use of nscd (Name Service Caching Daemon) in an IPA deployment. While nscd is extremely useful for reducing the load on the server, and for making clients more responsive, there are also drawbacks to its use.

nscd performs caching operations for all services that perform queries via the nsswitch interface, including getent. Because it performs both positive and negative caching, if it determines that a specific IPA user does not exist when a request is made, it marks this as a negative cache. Values stored in the cache remain until the cache expires, regardless of any changes that may have occurred on the server.

The results of such caching is that new users and memberships may not be visible, and users and memberships that have been removed may still be visible.

To alleviate the effects of this, you can avoid the use of nscd altogether, or change the default settings to use a shorter cache time. In particular, consider changing the following values in the /etc/nscd.conf file to suit the usage patterns of your deployment:

positive-time-to-live   group           3600
negative-time-to-live   group           60
positive-time-to-live   hosts           3600
negative-time-to-live   hosts           20

DNS and Kerberos

Kerberos, too, has very specific DNS requirements. Your Kerberos server requires a proper DNS A record, and reverse DNS needs to work correctly. Do not use CNAME or DDNS names, as it can cause major problems later. The IPA installation process includes checks to ensure that the IPA server name is a DNS A record and that its reverse matches its forward address.

Refer to IPA and DNS for more information on how these technologies work together.

Configuring /etc/hosts

You need to ensure that your /etc/hosts file is configured correctly, or the ipa-* commands may not work correctly.

The /etc/hosts file should list the FQDN for your IPA server before any aliases. You should also ensure that the hostname is not part of the localhost entry. The following is an example of a valid hosts file:

127.0.0.1       localhost.localdomain   localhost
::1     localhost6.localdomain6 localhost6
192.168.1.1     ipaserver.example.com      ipaserver

Hardware Requirements

The following table contains guidelines for Directory Server disk space and memory requirements based on on the number of entries that your organization requires. The values shown here assume that the entries in the LDIF file are approximately 100 bytes each and that only the recommended indices are configurable.

The system requirements for both 32-bit and 64-bit platforms are the same.

Operating system hardware requirements for IPA 1.0 Server
Criteria Requirements
Operating System Red Hat Enterprise Linux 5.1 Server or later, with the latest patches and upgrades

Fedora 7 or later, with the latest patches and upgrades

CPU Type Pentium 3 or higher; 500MHz or higher
Required Memory
Entries RAM
10,000 - 250,000 256 MB minimum
250,000 - 1,000,000 512 MB minimum
1,000,000 + 1 GB minimum
Hard Disk
Entries Disk Space
10,000 - 250,000 2GB
250,000 - 1,000,000 4GB
1,000,000 + 8GB

Installing the IPA Server Packages

This document distinguishes between commands to be run as root versus a regular user. Commands to be run as root are prefixed with a # symbol. Commands to be run as a regular user are prefixed with a $ symbol.

Note.pngNote:

Before starting the freeIPA installation, ensure that you update your system with all the latest packages.
If you are installing on 64-bit Red Hat Enterprise Linux 5.1, you need to update the krb5libs package before you install the ipa-server package.

1. freeIPA is currently only in the Fedora 7 and 8 updates-testing repository. It is in the regular repository for rawhide (Fedora 9). To install freeIPA you will need to enable the updates-testing repository. You can do this either by editing the /etc/yum.repos.d/updates-testing.repo file, or on the command line, as shown in step 2:

2. Run the following command to install the IPA server packages:

# yum install --enablerepo=updates-testing ipa-server

For Fedora 9 you do not need to include --enablerepo=updates-testing

This will install a large number of dependencies, including TurboGears, fedora-ds-base and krb5-server. Approximately 40 dependencies are required, depending on what is already installed.

3. freeIPA no longer requires a special mod_auth_kerb package but it does require a specific version of krb5-libs which contains a fix for Kerberos ticket delegation (look for "spnego" in the Changelog). The dependency should be handled by the ipa-server rpm, but if you want to verify manually, you need:

  • Fedora 7: 1.6.1-7 or higher
  • Fedora 8: 1.6.2-11 or higher
  • Fedora 9: all versions should be ok

If you previously installed the IPA-specific mod_auth_kerb package, you can remove it and replace it with the default Fedora version and ensure you have the minimum version of krb5-libs as listed above.


Note.pngNote:

If you are installing on Fedora 9, it is strongly recommend that you not use NetworkManager. Instead, run the following commands to use network to manage the network service:
# chkconfig NetworkManager off
# chkconfig network on
# service network start

Now you are ready to configure your IPA server.

Configuring the IPA Server

Use the ipa-server-install command to install the IPA server, which includes:

  • Configuring the Network Time Daemon (ntpd)
  • Creating and configuring an instance of Directory Server
  • Creating and configuring a Kerberos Key Distribution Center (krb5kdc)
  • Configuring Apache (httpd)
  • Configuring TurboGears
  • Updating the SELinux targeted policy

You can install the server interactively by running the command with no options, or by passing options directly to the ipa-server-install command. To view the available command-line options, run /usr/sbin/ipa-server-install --help

Note.pngNote:

If you are running IPA as a virtualized guest, you should not run the ntp daemon. In this case, you should pass the -N (no ntp) option to the ipa-server-install command.

To install the freeIPA server interactively:

1. Run the following command:

# ipa-server-install

2. When prompted, enter the server host name, realm name and other details.

The installation script compares the hostname returned by DNS to the hostname found in the /etc/hosts file. If the non-fully--qualified domain name appears first, the script aborts.

Note.pngNote:

The hostname that you enter into the ipa-server-install script must be the same as that returned by the hostname command, otherwise the Directory Server cannot use its own keytab. This results in commands such as ipa-finduser to fail.

3. Wait until the configuration script completes. Note that it can take several minutes to set up and configure all of the freeIPA requirements.

4. When the configuration script has completed, you should either reboot the server or at least restart the ssh server so that the Name Server Switch (nss) configuration is read when the service restarts.

To restart the ssh service, run the following command (existing connections are not terminated):
# service sshd restart

You can now proceed to test the configuration.

Testing the Configuration

The following examples assume that you are using EXAMPLE.COM as your realm.

Note.png Note:

The realm is used as the base DN in the directory instance, so it will be dc=example,dc=com.

When the installation is complete, all of the services should be running. You can test your installation as follows:

1. Use the kinit command to request a Kerberos ticket:

 $ kinit admin
 Password for admin@EXAMPLE.COM:

2. Use the klist command to display the list of Kerberos tickets:

 $ klist
 Ticket cache: FILE:/tmp/krb5cc_0
 Default principal: admin@EXAMPLE.COM
 Valid starting     Expires            Service principal
 03/05/08 02:47:53  03/06/08 02:47:50  krbtgt/EXAMPLE.COM@EXAMPLE.COM
 Kerberos 4 ticket cache: /tmp/tkt0
 klist: You have no tickets cached

3. Use the ipa user-find command to search for the admin user:

 $ ipa-finduser admin
 cn: Administrator
 homedirectory: /home/admin
 loginshell: /bin/bash
 uid: admin

If you receive output similar to the following, ensure that DNS is configured correctly:

Could not initialize GSSAPI: Unspecified GSS failure.
Minor code may provide more information/Server not found in Kerberos database.

Configuring your Browser

Firefox can use your Kerberos credentials for authentication, but you need to specify which domains you want to communicate with, and using which attributes.

1. Open Firefox, and type "about:config" in the Address Bar.

2. In the Search field, type "negotiate".

3. Ensure the following lines reflect your setup. Replace "example.com" with your own IPA server's domain, with a preceding period (.):

 network.negotiate-auth.trusted-uris  .example.com
 network.negotiate-auth.delegation-uris  .example.com
 network.negotiate-auth.using-native-gsslib true

4. In Firefox, navigate to your IPA server (use the fully-qualified domain name, for example, http://ipaserver.example.com). Ensure that there are no Kerberos authentication errors, and that you can see and interact with the Web interface.

Using a Browser on Another System

Use the following procedure to set up a browser on another system that already has Kerberos set up for a different realm.

1. Copy the /etc/krb5.conf file from the IPA server to the client system. Do not overwrite the existing krb5.conf file.

# scp /etc/krb5.conf root@ipaclient:/etc/krb5_ipa.conf

2. On the IPA client, open a terminal and run the following commands:

$ export KRB5_CONFIG=/etc/krb5_ipa.conf
$ kinit user@EXAMPLE.COM
$ /usr/bin/firefox

3. Configure the Firefox negotiate attributes as described in the Configuring your Browser section.

Now you should be able to connect to the IPA Web interface remotely.

Setting up Multi-Master Replication

Replication is the mechanism by which directory data is automatically copied from one Directory Server to another. Updates of any kind — adding, modifying, or deleting entries — are automatically mirrored to other Directory Servers using replication.

IPA 1.0 uses a number of scripts to install, configure, and manage replica servers and replication agreements. These are discussed in the following sections.

Preparing the Replica Servers

Replica servers require much the same preparation as IPA servers. That is, there should be no existing Directory Server installations, the ports required by IPA must be free and available, and the server's machine name be set and resolve to its public IP address (not to localhost). The replica server must also be able to contact the master LDAP server, so DNS or a similar lookup system must be working correctly.

Refer to Introduction to Installing IPA for more information about these and other considerations for installing an IPA server.

Installing the Server Packages

Follow the steps in Installing the IPA Server Packages to install all of the required packages for the replica server.

Note.pngNote:

Do not run the ipa-server-install script on the replica servers.

Creating the Replica Information File

You need to create a replica information file for each replica that you intend to create. This file contains various realm information required to correctly configure the replica server.

Before you create the replica information file, you need to ensure that the master IPA server is correctly configured and functioning properly. The master IPA server is the server from which all IPA replica servers will be created.

To create the replica information file:

Run the following command on the master IPA server, where ipareplica.example.com is the FQDN of the server where you are going to create the replica:

# ipa-replica-prepare ipareplica.example.com

This should produce output similar to the following:

Determining current realm name
Getting domain name from LDAP
Preparing replica for ipareplica.example.com from ipaserver.example.com
Creating SSL certificate for the Directory Server
Creating SSL certificate for the Web Server
Copying additional files
Finalizing configuration
Packaging the replica into replica-info-ipareplica.example.com

Note.png Note:

Each replica information file is created in the /var/lib/ipa/ directory, and named specifically for the replica server for which it is intended. You cannot use the same replica information file for multiple replicas.

Configuring an IPA Replica

After you have created the replica information file, you need to copy it to the replica server and run the required script to configure the replica.

To configure an IPA replica:

1. Copy the replica information file to the replica server:

# scp /var/lib/ipa/replica-info-ipareplica.example.com root@ipareplica:/var/lib/ipa/

2. On the replica server, run the replica installation script, passing it the replica information file you copied from the master:

# ipa-replica-install /var/lib/ipa/replica-info-ipareplica.example.com
The replica installation script runs a test to ensure that the replica file being installed matches the current hostname. If they do not match, the script returns a warning message and asks for confirmation. This could occur on a multi-homed machine, for example, where mismatched hostnames may not be an issue.

3. Enter the Directory Manager (DM) password when prompted.

The script then configures a directory server instance based on information in the replica information file, and initiates a replication process. When this has successfully completed, the script continues to set up a complete master replica of the IPA server.

Note.png Note:

You can only have a single Directory Server instance on an IPA server, the one used by IPA itself. If the replica installation script detects an existing Directory Server instance, you will be prompted to remove it.

Updating DNS for IPA Replicas

After you have configured a new IPA replica, you should update your DNS entries so that IPA clients can discover the new server. For example, for an IPA replica with a server name of $HOST, you should add the following entries to your zone file:

_ldap._tcp             IN SRV 0 100 389	$HOST
_kerberos._tcp         IN SRV 0 100 88 $HOST
_kerberos._udp         IN SRV 0 100 88 $HOST
_kerberos-master._tcp  IN SRV 0 100 88 $HOST
_kerberos-master._udp  IN SRV 0 100 88 $HOST
_kpasswd._tcp          IN SRV 0 100 464 $HOST
_kpasswd._udp          IN SRV 0 100 464 $HOST
_ntp._udp              IN SRV 0 100 123 $HOST

Managing Multi-Master Replication

You can use the ipa-replica-manage command to manage certain aspects of replication between IPA servers. This includes listing, adding, and deleting replication agreements, and also performing manual replication initialization and updates.

Initialization is typically only required when you first set up replication, or if a problem arises that causes replication to fail. Initialization erases all data on the target replica and re-copies all data from the master. That is, it completely destroys the database on the consumer and rebuilds it with data from the master.

Sending updates is the regular incremental replication protocol. Typically, this is not needed because the server sends changes when required, provided that the replication agreement schedule allows it.

Refer to the ipa-replica-manage man page for a full description of the available options.

Note.png Note:

There is no WebUI facility for managing IPA replicas. You need to use the command line.

Refer to the Managing Replication section of the Red Hat Directory Server Administration Guide for information about managing replication.

Troubleshooting Multi-Master Replication

Refer to the following sections in the Red Hat Director Server Administration Guide for information about troubleshooting replication:

Running IPA in a Virtual Host

If you have a standard Apache instance running on port 80, you can configure IPA to run on a secondary port, for example 8089. You should be aware, however, that running IPA in this configuration does not use SSL; all requests will go over standard HTTP.

Converting Your IPA Configuration to run as a VirtualHost

The following procedure assumes you already have IPA configured to run on port 80, and wish to move it to a different port.

1. Log in as the root user.

2. Edit the /etc/httpd/conf.d/ipa.conf file. Add the following three lines to the top of the file:

Listen 8089
NameVirtualHost *:8089
<VirtualHost *:8089>

3. Add the following line to the end of the file:

</VirtualHost>

This wraps the entire IPA configuration in a VirtualHost, and ensures that Apache is listening to that port.

Note.pngNote:

You can not use port 8080. This port is used by the ipa_webgui service.

4. Comment out the following rewrite rules from the /etc/httpd/conf.d/ipa.conf file:

----------------------------------------------------------------------
# Redirect to the fully-qualified hostname. Not redirecting to secure
# port so configuration files can be retrieved without requiring SSL.
RewriteCond %{HTTP_HOST}    !^host.foo.com$ [NC]
RewriteRule ^/(.*)          http://host.foo.com/$1 [L,R=301]

# Redirect to the secure port if not displaying an error or retrieving
# configuration.
RewriteCond %{SERVER_PORT}  !^443$
RewriteCond %{REQUEST_URI}  !^/(errors|config|favicon.ico)
RewriteRule ^/(.*)          https://host.foo.com/$1 [L,R=301,NC]
---------------------------------------------------------------------

5. Reload the httpd service.

# service httpd reload

IPA should now be running on port 8089, leaving port 8080 free for your normal web site.