ldap_id_mapping.xml revision 67ca9e7c006d8619f446c018eabf29eab1368ba5
<refsect1 id='idmap'>
<title>ID MAPPING</title>
<para>
The ID-mapping feature allows SSSD to act as a client of Active
Directory without requiring administrators to extend user attributes
to support POSIX attributes for user and group identifiers.
</para>
<para>
NOTE: When ID-mapping is enabled, the uidNumber and gidNumber
attributes are ignored. This is to avoid the possibility of conflicts
between automatically-assigned and manually-assigned values. If you
need to use manually-assigned values, ALL values must be
manually-assigned.
</para>
<refsect2 id='idmap_algorithm'>
<title>Mapping Algorithm</title>
<para>
Active Directory provides an objectSID for every user and group
object in the directory. This objectSID can be broken up into
components that represent the Active Directory domain identity and
the relative identifier (RID) of the user or group object.
</para>
<para>
The SSSD ID-mapping algorithm takes a range of available UIDs and
divides it into equally-sized component sections - called
"slices"-. Each slice represents the space available to an Active
Directory domain.
</para>
<para>
When a user or group entry for a particular domain is encountered
for the first time, the SSSD allocates one of the available slices
for that domain. In order to make this slice-assignment repeatable
on different client machines, we select the slice based on the
following algorithm:
</para>
<para>
The SID string is passed through the murmurhash3 algorithm to
convert it to a 32-bit hashed value. We then take the modulus of
this value with the total number of available slices to pick the
slice.
</para>
<para>
NOTE: It is possible to encounter collisions in the hash and
subsequent modulus. In these situations, we will select the next
available slice, but it may not be possible to reproduce the same
exact set of slices on other machines (since the order that they
are encountered will determine their slice). In this situation, it
is recommended to either switch to using explicit POSIX attributes
in Active Directory (disabling ID-mapping) or configure a default
domain to guarantee that at least one is always consistent. See
<quote>Configuration</quote> for details.
</para>
</refsect2>
<refsect2 id='idmap_config'>
<title>Configuration</title>
<para>
section):
</para>
<para>
<programlisting>
ldap_id_mapping = True
ldap_schema = ad
</programlisting>
</para>
<para>
The default configuration results in configuring 10,000 slices,
each capable of holding up to 200,000 IDs, starting from 10,001
and going up to 2,000,100,000. This should be sufficient for
most deployments.
</para>
<refsect3 id='idmap_advanced_config'>
<title>Advanced Configuration</title>
<variablelist>
<varlistentry>
<term>ldap_idmap_range_min (integer)</term>
<listitem>
<para>
Specifies the lower bound of the range of POSIX IDs to
use for mapping Active Directory user and group SIDs.
</para>
<para>
NOTE: This option is different from
<quote>min_id</quote> in that <quote>min_id</quote>
acts to filter the output of requests to this domain,
whereas this option controls the range of ID
assignment. This is a subtle distinction, but the
good general advice would be to have
<quote>min_id</quote> be less-than or equal to
<quote>ldap_idmap_range_min</quote>
</para>
<para>
Default: 200000
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldap_idmap_range_max (integer)</term>
<listitem>
<para>
Specifies the upper bound of the range of POSIX IDs to
use for mapping Active Directory user and group SIDs.
</para>
<para>
NOTE: This option is different from
<quote>max_id</quote> in that <quote>max_id</quote>
acts to filter the output of requests to this domain,
whereas this option controls the range of ID
assignment. This is a subtle distinction, but the
good general advice would be to have
<quote>max_id</quote> be greater-than or equal to
<quote>ldap_idmap_range_max</quote>
</para>
<para>
Default: 2000200000
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldap_idmap_range_size (integer)</term>
<listitem>
<para>
Specifies the number of IDs available for each slice.
If the range size does not divide evenly into the min
and max values, it will create as many complete slices
as it can.
</para>
<para>
Default: 200000
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldap_idmap_default_domain_sid (string)</term>
<listitem>
<para>
Specify the domain SID of the default domain. This
will guarantee that this domain will always be
assigned to slice zero in the ID map, bypassing
the murmurhash algorithm described above.
</para>
<para>
Default: not set
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldap_idmap_default_domain (string)</term>
<listitem>
<para>
Specify the name of the default domain.
</para>
<para>
Default: not set
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>ldap_idmap_autorid_compat (boolean)</term>
<listitem>
<para>
Changes the behavior of the ID-mapping algorithm
to behave more similarly to winbind's
<quote>idmap_autorid</quote> algorithm.
</para>
<para>
When this option is configured, domains will be
allocated starting with slice zero and increasing
monatomically with each additional domain.
</para>
<para>
NOTE: This algorithm is non-deterministic (it
depends on the order that users and groups are
requested). If this mode is required for
compatibility with machines running winbind, it
is recommended to also use the
<quote>ldap_idmap_default_domain_sid</quote>
option to guarantee that at least one domain is
consistently allocated to slice zero.
</para>
<para>
Default: False
</para>
</listitem>
</varlistentry>
</variablelist>
</refsect3>
</refsect2>
</refsect1>