Section: MIT Kerberos (7)
Page Index


kerberos - Overview of using Kerberos  


The Kerberos system authenticates individual users in a network environment. After authenticating yourself to Kerberos, you can use Kerberos-enabled programs without having to present passwords or certificates to those programs.

If you receive the following response from kinit(1):

kinit: Client not found in Kerberos database while getting initial credentials

you haven't been registered as a Kerberos user. See your system administrator.

A Kerberos name usually contains three parts. The first is the primary, which is usually a user's or service's name. The second is the instance, which in the case of a user is usually null. Some users may have privileged instances, however, such as root or admin. In the case of a service, the instance is the fully qualified name of the machine on which it runs; i.e. there can be an ssh service running on the machine ABC (ssh/ABC@REALM), which is different from the ssh service running on the machine XYZ (ssh/XYZ@REALM). The third part of a Kerberos name is the realm. The realm corresponds to the Kerberos service providing authentication for the principal. Realms are conventionally all-uppercase, and often match the end of hostnames in the realm (for instance, might be in realm EXAMPLE.COM).

When writing a Kerberos name, the principal name is separated from the instance (if not null) by a slash, and the realm (if not the local realm) follows, preceded by an "@" sign. The following are examples of valid Kerberos names:


When you authenticate yourself with Kerberos you get an initial Kerberos ticket. (A Kerberos ticket is an encrypted protocol message that provides authentication.) Kerberos uses this ticket for network utilities such as ssh. The ticket transactions are done transparently, so you don't have to worry about their management.

Note, however, that tickets expire. Administrators may configure more privileged tickets, such as those with service or instance of root or admin, to expire in a few minutes, while tickets that carry more ordinary privileges may be good for several hours or a day. If your login session extends beyond the time limit, you will have to re-authenticate yourself to Kerberos to get new tickets using the kinit(1) command.

Some tickets are renewable beyond their initial lifetime. This means that kinit -R can extend their lifetime without requiring you to re-authenticate.

If you wish to delete your local tickets, use the kdestroy(1) command.

Kerberos tickets can be forwarded. In order to forward tickets, you must request forwardable tickets when you kinit. Once you have forwardable tickets, most Kerberos programs have a command line option to forward them to the remote host. This can be useful for, e.g., running kinit on your local machine and then sshing into another to do work. Note that this should not be done on untrusted machines since they will then have your tickets.  


Several environment variables affect the operation of Kerberos-enabled programs. These include:

Default name for the credentials cache file, in the form TYPE:residual. The type of the default cache may determine the availability of a cache collection. FILE is not a collection type; KEYRING, DIR, and KCM are.

If not set, the value of default_ccache_name from configuration files (see KRB5_CONFIG) will be used. If that is also not set, the default type is FILE, and the residual is the path /tmp/krb5cc_*uid*, where uid is the decimal user ID of the user.

Specifies the location of the default keytab file, in the form TYPE:residual. If no type is present, the FILE type is assumed and residual is the pathname of the keytab file. If unset, FILE:/etc/krb5.keytab will be used.
Specifies the location of the Kerberos configuration file. The default is /etc/krb5.conf. Multiple filenames can be specified, separated by a colon; all files which are present will be read.
Specifies the location of the KDC configuration file, which contains additional configuration directives for the Key Distribution Center daemon and associated programs. The default is /var/kerberos/krb5kdc/kdc.conf.
Specifies the default type of replay cache to use for servers. Valid types include dfl for the normal file type and none for no replay cache. The default is dfl.
Specifies the default directory for replay caches used by servers. The default is the value of the TMPDIR environment variable, or /var/tmp if TMPDIR is not set.
Specifies a filename to write trace log output to. Trace logs can help illuminate decisions made internally by the Kerberos libraries. For example, env KRB5_TRACE=/dev/stderr kinit would send tracing information for kinit(1) to /dev/stderr. The default is not to write trace log output anywhere.
Default client keytab file name. If unset, FILE:/var/kerberos/krb5/user/%{euid}/client.keytab will be used).
kprop(8) port to use. Defaults to 754.

Most environment variables are disabled for certain programs, such as login system programs and setuid programs, which are designed to be secure when run within an untrusted process environment.



kdestroy(1), kinit(1), klist(1), kswitch(1), kpasswd(1), ksu(1), krb5.conf(5), kdc.conf(5), kadmin(1), kadmind(8), kdb5_util(8), krb5kdc(8)  




Steve Miller, MIT Project Athena/Digital Equipment Corporation
Clifford Neuman, MIT Project Athena
Greg Hudson, MIT Kerberos Consortium
Robbie Harwood, Red Hat, Inc.



The MIT Kerberos 5 implementation was developed at MIT, with contributions from many outside parties. It is currently maintained by the MIT Kerberos Consortium.  


Copyright 1985, 1986, 1989-1996, 2002, 2011, 2018 Masachusetts Institute of Technology  




1985-2019, MIT