6 Network Authentication with Kerberos #

Your first contact with Kerberos is quite similar to any login procedure at a normal networking system. Enter your user name. This piece of information and the name of the ticket-granting service are sent to the authentication server (Kerberos). If the authentication server knows you, it generates a random session key for further use between your client and the ticket-granting server. Now the authentication server prepares a ticket for the ticket-granting server. The ticket contains the following information—all encrypted with a session key only the authentication server and the ticket-granting server know:

This ticket is then sent back to the client together with the session key, again in encrypted form, but this time the private key of the client is used. This private key is only known to Kerberos and the client, because it is derived from your user password. Now that the client has received this response, you are prompted for your password. This password is converted into the key that can decrypt the package sent by the authentication server. The package is “unwrapped” and password and key are erased from the workstation's memory. As long as the lifetime given to the ticket used to obtain other tickets does not expire, your workstation can prove your identity.

To request a service from any server in the network, the client application needs to prove its identity to the server. Therefore, the application generates an authenticator. An authenticator consists of the following components:

All this information is encrypted using the session key that the client has already received for this special server. The authenticator and the ticket for the server are sent to the server. The server uses its copy of the session key to decrypt the authenticator, which gives it all the information needed about the client requesting its service, to compare it to that contained in the ticket. The server checks if the ticket and the authenticator originate from the same client.

Without any security measures implemented on the server side, this stage of the process would be an ideal target for replay attacks. Someone could try to resend a request stolen off the net some time before. To prevent this, the server does not accept any request with a time stamp and ticket received previously. In addition to that, a request with a time stamp differing too much from the time the request is received is ignored.

Kerberos authentication can be used in both directions. It is not only a question of the client being the one it claims to be. The server should also be able to authenticate itself to the client requesting its service. Therefore, it sends an authenticator itself. It adds one to the checksum it received in the client's authenticator and encrypts it with the session key, which is shared between it and the client. The client takes this response as a proof of the server's authenticity and they both start cooperating.

Tickets are designed to be used for one server at a time. Therefore, you need to get a new ticket each time you request another service. Kerberos implements a mechanism to obtain tickets for individual servers. This service is called the “ticket-granting service”. The ticket-granting service is a service (like any other service mentioned before) and uses the same access protocols that have already been outlined. Any time an application needs a ticket that has not already been requested, it contacts the ticket-granting server. This request consists of the following components:

Like any other server, the ticket-granting server now checks the ticket-granting ticket and the authenticator. If they are considered valid, the ticket-granting server builds a new session key to be used between the original client and the new server. Then the ticket for the new server is built, containing the following information:

The new ticket has a lifetime, which is either the remaining lifetime of the ticket-granting ticket or the default for the service. The lesser of both values is assigned. The client receives this ticket and the session key, which are sent by the ticket-granting service. But this time the answer is encrypted with the session key that came with the original ticket-granting ticket. The client can decrypt the response without requiring the user's password when a new service is contacted. Kerberos can thus acquire ticket after ticket for the client without bothering the user.

Any Kerberos environment must meet the following requirements to be fully functional:

The following figure depicts a simple example network with only the minimum components needed to build a Kerberos infrastructure. Depending on the size and topology of your deployment, your setup may vary.

Figure 6.1: Kerberos Network Topology #

 

Tip: Configuring Subnet Routing

For a setup similar to the one in Figure 6.1, “Kerberos Network Topology”, configure routing between the two subnets (192.168.1.0/24 and 192.168.2.0/24). Refer to 16.4.1.5項 「ルーティングの設定」 for more information on configuring routing with YaST.

The domain of a Kerberos installation is called a realm and is identified by a name, such as EXAMPLE.COM or simply ACCOUNTING. Kerberos is case-sensitive, so example.com is actually a different realm than EXAMPLE.COM. Use the case you prefer. It is common practice, however, to use uppercase realm names.

It is also a good idea to use your DNS domain name (or a subdomain, such as ACCOUNTING.EXAMPLE.COM). As shown below, your life as an administrator can be much easier if you configure your Kerberos clients to locate the KDC and other Kerberos services via DNS. To do so, it is helpful if your realm name is a subdomain of your DNS domain name.

Unlike the DNS name space, Kerberos is not hierarchical. So if you have a realm named EXAMPLE.COM with two “subrealms” named DEVELOPMENT and ACCOUNTING, these subordinate realms do not inherit principals from EXAMPLE.COM. Instead, you would have three separate realms, and you would need to configure cross-realm authentication for each realm, so that users from one realm can interact with servers or other users from another realm.

For the sake of simplicity, let us assume you are setting up just one realm for your entire organization. For the remainder of this section, the realm name EXAMPLE.COM is used in all examples.

The first thing required to use Kerberos is a machine that acts as the key distribution center, or KDC for short. This machine holds the entire Kerberos user database with passwords and all information.

The KDC is the most important part of your security infrastructure—if someone breaks into it, all user accounts and all of your infrastructure protected by Kerberos is compromised. An attacker with access to the Kerberos database can impersonate any principal in the database. Tighten security for this machine as much as possible:

To use Kerberos successfully, make sure that all system clocks within your organization are synchronized within a certain range. This is important because Kerberos protects against replayed credentials. An attacker might be able to observe Kerberos credentials on the network and reuse them to attack the server. Kerberos employs several defenses to prevent this. One of them is that it puts time stamps into its tickets. A server receiving a ticket with a time stamp that differs from the current time rejects the ticket.

Kerberos allows a certain leeway when comparing time stamps. However, computer clocks can be very inaccurate in keeping time—it is not unheard of for PC clocks to lose or gain half an hour during a week. For this reason, configure all hosts on the network to synchronize their clocks with a central time source.

A simple way to do so is by installing an NTP time server on one machine and having all clients synchronize their clocks with this server. Do this either by running an NTP daemon in client mode on all these machines or by running ntpdate once a day from all clients (this solution probably works for a few clients only). The KDC itself needs to be synchronized to the common time source as well. Because running an NTP daemon on this machine would be a security risk, it is probably a good idea to do this by running ntpdate via a cron job. To configure your machine as an NTP client, proceed as outlined in 25.1項 「YaSTでのNTPクライアントの設定」.

A different way to secure the time service and still use the NTP daemon is to attach a hardware reference clock to a dedicated NTP server and an additional hardware reference clock to the KDC.

It is also possible to adjust the maximum deviation Kerberos allows when checking time stamps. This value (called clock skew) can be set in the krb5.conf file as described in Section 6.5.6.3, “Adjusting the Clock Skew”.

This section covers the initial configuration and installation of the KDC, including the creation of an administrative principal. This procedure consists of several steps:

kdb5_util create -r EXAMPLE.COM -s

Initializing database '/var/lib/kerberos/krb5kdc/principal' for realm 'EXAMPLE.COM',
master key name 'K/M@EXAMPLE.COM'
You will be prompted for the database Master Password.
It is important that you NOT FORGET this password.
Enter KDC database master key:  1
Re-enter KDC database master key to verify:  2

kadmin.local

kadmin> listprincs

K/M@EXAMPLE.COM
kadmin/admin@EXAMPLE.COM
kadmin/changepw@EXAMPLE.COM
krbtgt/EXAMPLE.COM@EXAMPLE.COM

kadmin.local

kadmin> ank suzanne

suzanne@EXAMPLE.COM's Password: 1
Verifying password: 2

sudo systemctl start krb5kdc
sudo systemctl start kadmind

sudo systemctl enable krb5kdc kadmind

When the supporting infrastructure is in place (DNS, NTP) and the KDC has been properly configured and started, configure the client machines. To configure a Kerberos client, use one of the two manual approaches described below.

When configuring Kerberos, there are two approaches you can take—static configuration in the /etc/krb5.conf file or dynamic configuration with DNS. With DNS configuration, Kerberos applications try to locate the KDC services using DNS records. With static configuration, add the host names of your KDC server to krb5.conf (and update the file whenever you move the KDC or reconfigure your realm in other ways).

DNS-based configuration is generally a lot more flexible and the amount of configuration work per machine is a lot less. However, it requires that your realm name is either the same as your DNS domain or a subdomain of it. Configuring Kerberos via DNS also creates a security issue: an attacker can seriously disrupt your infrastructure through your DNS (by shooting down the name server, spoofing DNS records, etc.). However, this amounts to a denial of service at worst. A similar scenario applies to the static configuration case unless you enter IP addresses in krb5.conf instead of host names.

[libdefaults]
        default_realm = EXAMPLE.COM

[realms]
        EXAMPLE.COM = {
                kdc = kdc.example.com
                admin_server = kdc.example.com
        }

[domain_realm]
.example.com = EXAMPLE.COM
www.example.org = EXAMPLE.COM

_kerberos.www.example.org.  IN TXT "EXAMPLE.COM"

[libdefaults]
        clockskew = 60

To be able to add and remove principals from the Kerberos database without accessing the KDC's console directly, tell the Kerberos administration server which principals are allowed to do what by editing /var/lib/kerberos/krb5kdc/kadm5.acl. The ACL (access control list) file allows you to specify privileges with a precise degree of control. For details, refer to the manual page with man 8 kadmind.

For now, grant yourself the privilege to administer the database by putting the following line into the file:

suzanne/admin              *

Replace the user name suzanne with your own. Restart kadmind for the change to take effect.

You should now be able to perform Kerberos administration tasks remotely using the kadmin tool. First, obtain a ticket for your admin role and use that ticket when connecting to the kadmin server:

kadmin -p suzanne/admin
Authenticating as principal suzanne/admin@EXAMPLE.COM with password.
Password for suzanne/admin@EXAMPLE.COM:
kadmin:  getprivs
current privileges: GET ADD MODIFY DELETE
kadmin:

Using the getprivs command, verify which privileges you have. The list shown above is the full set of privileges.

As an example, modify the principal suzanne:

kadmin -p suzanne/admin
Authenticating as principal suzanne/admin@EXAMPLE.COM with password.
Password for suzanne/admin@EXAMPLE.COM:

kadmin:  getprinc suzanne
Principal: suzanne@EXAMPLE.COM
Expiration date: [never]
Last password change: Wed Jan 12 17:28:46 CET 2005
Password expiration date: [none]
Maximum ticket life: 0 days 10:00:00
Maximum renewable life: 7 days 00:00:00
Last modified: Wed Jan 12 17:47:17 CET 2005 (admin/admin@EXAMPLE.COM)
Last successful authentication: [never]
Last failed authentication: [never]
Failed password attempts: 0
Number of keys: 2
Key: vno 1, Triple DES cbc mode with HMAC/sha1, no salt
Key: vno 1, DES cbc mode with CRC-32, no salt
Attributes:
Policy: [none]

kadmin:  modify_principal -maxlife "8 hours" suzanne
Principal "suzanne@EXAMPLE.COM" modified.
kadmin:  getprinc suzanne
Principal: suzanne@EXAMPLE.COM
Expiration date: [never]
Last password change: Wed Jan 12 17:28:46 CET 2005
Password expiration date: [none]
Maximum ticket life: 0 days 08:00:00
Maximum renewable life: 7 days 00:00:00
Last modified: Wed Jan 12 17:59:49 CET 2005 (suzanne/admin@EXAMPLE.COM)
Last successful authentication: [never]
Last failed authentication: [never]
Failed password attempts: 0
Number of keys: 2
Key: vno 1, Triple DES cbc mode with HMAC/sha1, no salt
Key: vno 1, DES cbc mode with CRC-32, no salt
Attributes:
Policy: [none]
kadmin:

This changes the maximum ticket life time to eight hours. For more information about the kadmin command and the options available, see the krb5-doc package or refer to the man 8 kadmin manual page.

So far, only user credentials have been discussed. However, Kerberos-compatible services usually need to authenticate themselves to the client user, too. Therefore, special service principals must be in the Kerberos database for each service offered in the realm. For example, if ldap.example.com offers an LDAP service, you need a service principal, ldap/ldap.example.com@EXAMPLE.COM, to authenticate this service to all clients.

The naming convention for service principals is SERVICE/HOSTNAME@REALM, where HOSTNAME is the host's fully qualified host name.

Valid service descriptors are:

Service principals are similar to user principals, but have significant differences. The main difference between a user principal and a service principal is that the key of the former is protected by a password. When a user obtains a ticket-granting ticket from the KDC, they needs to type their password, so Kerberos can decrypt the ticket. It would be inconvenient for system administrators to obtain new tickets for the SSH daemon every eight hours or so.

Instead, the key required to decrypt the initial ticket for the service principal is extracted by the administrator from the KDC only once and stored in a local file called the keytab. Services such as the SSH daemon read this key and use it to obtain new tickets automatically, when needed. The default keytab file resides in /etc/krb5.keytab.

To create a host service principal for jupiter.example.com enter the following commands during your kadmin session:

kadmin -p suzanne/admin
Authenticating as principal suzanne/admin@EXAMPLE.COM with password.
Password for suzanne/admin@EXAMPLE.COM:
kadmin:  addprinc -randkey host/jupiter.example.com
WARNING: no policy specified for host/jupiter.example.com@EXAMPLE.COM;
defaulting to no policy
Principal "host/jupiter.example.com@EXAMPLE.COM" created.

Instead of setting a password for the new principal, the -randkey flag tells kadmin to generate a random key. This is used here because no user interaction is wanted for this principal. It is a server account for the machine.

Finally, extract the key and store it in the local keytab file /etc/krb5.keytab. This file is owned by the superuser, so you must be root to execute the next command in the kadmin shell:

kadmin:  ktadd host/jupiter.example.com
Entry for principal host/jupiter.example.com with kvno 3, encryption type Triple
DES cbc mode with HMAC/sha1 added to keytab WRFILE:/etc/krb5.keytab.
Entry for principal host/jupiter.example.com with kvno 3, encryption type DES
cbc mode with CRC-32 added to keytab WRFILE:/etc/krb5.keytab.
kadmin:

When completed, make sure that you destroy the admin ticket obtained with kinit above with kdestroy.

Warning: Incomplete Configuration Locks Users Out

An incomplete Kerberos configuration may completely lock you out of your system, including the root user. To prevent this, add the ignore_unknown_principals directive to the pam_krb5 module after you have added the pam_krb5 module to the existing PAM configuration files as described below.

tux > sudo pam-config --add --krb5-ignore_unknown_principals

This will direct the pam_krb5 module to ignore some errors that would otherwise cause the account phase to fail.

SUSE® Linux Enterprise Server comes with a PAM module named pam_krb5, which supports Kerberos login and password update. This module can be used by applications such as console login, su, and graphical login applications like GDM. That is, it can be used in all cases where the user enters a password and expects the authenticating application to obtain an initial Kerberos ticket on their behalf. To configure PAM support for Kerberos, use the following command:

tux > sudo pam-config --add --krb5

The above command adds the pam_krb5 module to the existing PAM configuration files and makes sure it is called in the right order. To make precise adjustments to the way in which pam_krb5 is used, edit the file /etc/krb5.conf and add default applications to PAM. For details, refer to the manual page with man 5 pam_krb5.

The pam_krb5 module was specifically not designed for network services that accept Kerberos tickets as part of user authentication. This is an entirely different matter, and is discussed below.

OpenSSH supports Kerberos authentication in both protocol version 1 and 2. In version 1, there are special protocol messages to transmit Kerberos tickets. Version 2 does not use Kerberos directly anymore, but relies on GSSAPI, the General Security Services API. This is a programming interface that is not specific to Kerberos—it was designed to hide the peculiarities of the underlying authentication system, be it Kerberos, a public-key authentication system like SPKM, or others. However, the included GSSAPI library only supports Kerberos.

To use sshd with Kerberos authentication, edit /etc/ssh/sshd_config and set the following options:

# These are for protocol version 1
#
# KerberosAuthentication yes
# KerberosTicketCleanup yes

# These are for version 2 - better to use this
GSSAPIAuthentication yes
GSSAPICleanupCredentials yes

Then restart your SSH daemon using sudo systemctl restart sshd.

To use Kerberos authentication with protocol version 2, enable it on the client side as well. Do this either in the systemwide configuration file /etc/ssh/ssh_config or on a per-user level by editing ~/.ssh/config. In both cases, add the option GSSAPIAuthentication yes.

You should now be able to connect using Kerberos authentication. Use klist to verify that you have a valid ticket, then connect to the SSH server. To force SSH protocol version 1, specify the -1 option on the command line.

Tip: Additional Information

The file /usr/share/doc/packages/openssh/README.kerberos discusses the interaction of OpenSSH and Kerberos in more detail.

Tip: Additional Directives for Protocol Version 2

The GSSAPIKeyExchange mechanism (RFC 4462) is supported. This directive specifies how host keys are exchanged. For more information, see the sshd_config manual page (man sshd_config).

When using Kerberos, one way to distribute the user information (such as user ID, groups, and home directory) in your local network is to use LDAP. This requires a strong authentication mechanism that prevents packet spoofing and other attacks. One solution is to use Kerberos for LDAP communication, too.

OpenLDAP implements most authentication flavors through SASL, the simple authentication session layer. SASL is a network protocol designed for authentication. The SASL implementation is cyrus-sasl, which supports several authentication flavors. Kerberos authentication is performed through GSSAPI (General Security Services API). By default, the SASL plug-in for GSSAPI is not installed. Install the cyrus-sasl-gssapi with YaST.

To enable Kerberos to bind to the OpenLDAP server, create a principal ldap/ldap.example.com and add that to the keytab.

By default, the LDAP server slapd runs as user and group ldap, while the keytab file is readable by root only. Therefore, either change the LDAP configuration so the server runs as root or make the keytab file readable by the group ldap. The latter is done automatically by the OpenLDAP start script (/usr/lib/openldap/start) if the keytab file has been specified in the OPENLDAP_KRB5_KEYTAB variable in /etc/sysconfig/openldap and the OPENLDAP_CHOWN_DIRS variable is set to yes, which is the default setting. If OPENLDAP_KRB5_KEYTAB is left empty, the default keytab under /etc/krb5.keytab is used and you must adjust the privileges yourself as described below.

To run slapd as root, edit /etc/sysconfig/openldap. Disable the OPENLDAP_USER and OPENLDAP_GROUP variables by putting a comment character in front of them.

To make the keytab file readable by group LDAP, execute

chgrp ldap /etc/krb5.keytab
chmod 640 /etc/krb5.keytab

A third (and maybe the best) solution is to tell OpenLDAP to use a special keytab file. To do this, start kadmin, and enter the following command after you have added the principal ldap/ldap.example.com:

tux > sudo ktadd -k /etc/openldap/ldap.keytab ldap/ldap.example.com@EXAMPLE.COM

Then in the shell run:

chown ldap.ldap /etc/openldap/ldap.keytab
chmod 600 /etc/openldap/ldap.keytab

To tell OpenLDAP to use a different keytab file, change the following variable in /etc/sysconfig/openldap:

OPENLDAP_KRB5_KEYTAB="/etc/openldap/ldap.keytab"

Finally, restart the LDAP server using sudo systemctl restart slapd.

ldapsearch -b ou=people,dc=example,dc=com '(uid=suzanne)'

SASL/GSSAPI authentication started
SASL SSF: 56
SASL installing layers
[...]

# suzanne, people, example.com
dn: uid=suzanne,ou=people,dc=example,dc=com
uid: suzanne
cn: Suzanne Geeko
[...]

dn: cn=config
add: olcAuthzRegexp
olcAuthzRegexp: uid=(.*),cn=GSSAPI,cn=auth uid=$1,ou=people,dc=example,dc=com

The one behavioral difference between sec=sys and the various Kerberos security levels that might be visible is related to group membership. In Unix and Linux, each file system access comes from a process that is owned by a particular user and has a particular group owner and a number of supplemental groups. Access rights to files can vary based on the owner and the various groups.

With sec=sys, the user-id, group-id, and a list of up to 16 supplemental groups are sent to the server in each request.

If a user is a member of more than 16 supplemental groups, the extra groups are lost and some files may not be accessible over NFS that the user would normally expect to have access to. For this reason, most sites that use NFS find a way to limit all users to at most 16 supplemental groups.

If the user runs the newgrp command or runs a set-group-id program, either of which can change the list of groups they are a member of, these changes take effect immediately and provide different accesses over NFS.

With Kerberos, group information is not sent in requests at all. Only the user is identified (using a Kerberos "Principal"), and the server performs a lookup to determine the user ID and group list for that principal. This means that if the user is a member of more than 16 groups, all of these group memberships will be used in determining file access permissions. However it also means that if the user changes a group-id on the client in some way, the server will not notice the change and will not take that into account in determining access rights.

Is most cases, the improvement of having access to more groups brings a real benefit, and the loss of not being able to change groups is not noticed as it is not widely used. A site administrator considering the use of Kerberos should be aware of the difference though and ensure that it will not actually cause problems.

Using Kerberos for security requires extra CPU power for encrypting and decrypting messages. How much extra CPU power is required and whether the difference is noticeable will vary with different hardware and different applications. If the server or client are already saturating the available CPU power, it is likely that a performance drop will be measurable when switching from sec=sys to Kerberos. If there is spare CPU capacity available, it is quite possible that the transition will not result in any throughput change. The only way to be sure how much impact the use of Kerberos will have is to test your load on your hardware.

The only configuration options that might reduce the load will also reduce the quality of the protection offered. sec=krb5 should produce noticeably less load than sec=krb5p but, as discussed above, it doesn't produce very strong security. Similarly it is possible to adjust the list of cyphers that Kerberos can choose from, and this might change the CPU requirement. However the defaults are carefully chosen and should not be changed without similar careful consideration.

The other possible performance issue when configuring NFS to use Kerberos involves availability of the Kerberos authentication servers, known as the KDC or Key Distribution Center.

The use of NFS adds load to such servers to the same degree that adding the use of Kerberos for any other services adds some load. Every time a given user (Kerberos principal) establishes a session with a service, for example by accessing files exported by a particular NFS server, the client needs to negotiate with the KDC. Once a session key has been negotiated, the client server can communicate without further help for many hours, depending on details of the Kerberos configuration, particularly the ticket_lifetime setting.

The concerns most likely to affect the provisioning of Kerberos KDC servers are availability and peak usage.

As with other core services such as DNS, LDAP or similar name-lookup services, having two servers that are reasonably "close" to every client provides good availability for modest resources. Kerberos allows for multiple KDC servers with flexible models for database propagation, so distributing servers as needed around campuses, buildings, and even cabinets is fairly straight forward. The best mechanism to ensure each client finds a nearby Kerberos server is to use split-horizon DNS with each building (or similar) getting different details from the DNS server. If this is not possible, then managing the /etc/krb5.conf file to be different at different locations is a suitable alternative.

As access to the Kerberos KDC is infrequent, load is only likely to be a problem at peak times. If thousands of people all log in between 9:00 and 9:05, then the servers will receive many more requests-per-minute than they might in the middle of the night. The load on the Kerberos server is likely to be more than that on an LDAP server, but not orders of magnitude more. A sensible guideline is to provision Kerberos replicas in the same manner that you provision LDAP replicas, and then monitor performance to determine if demand ever exceeds capacity.

One service of the Kerberos KDC that is not easily distributed is the handling of updates, such as password changes and new user creation. These must happen at a single master KDC.

These updates are not likely to happen with such frequency that any significant load will be generated, but availability could be an issue. It can be annoying if you want to create a new user or change a password, and the master KDC on the other side of the world is temporarily unavailable.

When an organization is geographically distributed and has a policy of handling administration tasks locally at each site, it can be beneficial to create multiple Kerberos domains, one for each administrative center. Each domain would then have its own master KDC which would be geographically local. Users in one domain can still get access to resources in another domain by setting up trust relationships between domains.

The easiest arrangement for multiple domains is to have a global domain (such as EXAMPLE.COM and various local domains (such as ASIA.EXAMPLE.COM, EUROPE.EXAMPLE.COM, etc). If the global domain is configured to trust each local domain, and each local domain is configured to trust the global domain, then fully transitive trust is available between any pair of domains, and any principal can establish a secure connection with any service. Ensuring appropriate access rights to resources, for example files, provided by that service will be dependent on the user name lookup service used, and the functionality of the NFS file server, and is beyond the scope of this document.