1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
|
<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>
<para>
Please note that changing the ID mapping related configuration options will
cause user and group IDs to change. At the moment, SSSD does not support
changing IDs, so the SSSD database must be removed. Because cached passwords
are also stored in the database, removing the database should only be
performed while the authentication servers are reachable, otherwise users
might get locked out. In order to cache the password, an authentication must
be performed. It is not sufficient to use <citerefentry>
<refentrytitle>sss_cache</refentrytitle> <manvolnum>8</manvolnum>
</citerefentry> to remove the database, rather the process consists of:
<itemizedlist>
<listitem>
<para>
Making sure the remote servers are reachable
</para>
</listitem>
<listitem>
<para>
Stopping the SSSD service
</para>
</listitem>
<listitem>
<para>
Removing the database
</para>
</listitem>
<listitem>
<para>
Starting the SSSD service
</para>
</listitem>
</itemizedlist>
Moreover, as the change of IDs might necessitate the adjustment of other
system properties such as file and directory ownership, it's advisable to
plan ahead and test the ID mapping configuration thoroughly.
</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>
Minimum configuration (in the <quote>[domain/DOMAINNAME]</quote> 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 200,000 and going up to
2,000,200,000. 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 (inclusive) bound of the range of POSIX IDs to use for
mapping Active Directory user and group SIDs. It is the first POSIX ID which
can be used for the mapping.
</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 (exclusive) bound of the range of POSIX IDs to use for
mapping Active Directory user and group SIDs. It is the first POSIX ID which
cannot be used for the mapping anymore, i.e. one larger than the last one
which can be used for the mapping.
</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>
NOTE: The value of this option must be at least as large as the highest user
RID planned for use on the Active Directory server. User lookups and login
will fail for any user whose RID is greater than this value.
</para>
<para>
For example, if your most recently-added Active Directory user has
objectSid=S-1-5-21-2153326666-2176343378-3404031434-1107,
<quote>ldap_idmap_range_size</quote> must be at least 1108 as range size is
equal to maximal SID minus minimal SID plus one (e.g. 1108 = 1107 - 0 + 1).
</para>
<para>
It is important to plan ahead for future expansion, as changing this value
will result in changing all of the ID mappings on the system, leading to
users with different local IDs than they previously had.
</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 monotonically 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>
<varlistentry>
<term>ldap_idmap_helper_table_size (integer)</term>
<listitem>
<para>
Maximal number of secondary slices that is tried when performing mapping
from UNIX id to SID.
</para>
<para>
Note: Additional secondary slices might be generated when SID is being
mapped to UNIX id and RID part of SID is out of range for secondary slices
generated so far. If value of ldap_idmap_helper_table_size is equal to 0
then no additional secondary slices are generated.
</para>
<para>
Default: 10
</para>
</listitem>
</varlistentry>
</variablelist>
</refsect3>
</refsect2>
<refsect2 id='well_known_sids'>
<title>Well-Known SIDs</title>
<para>
SSSD supports to look up the names of Well-Known SIDs, i.e. SIDs with a
special hardcoded meaning. Since the generic users and groups related to
those Well-Known SIDs have no equivalent in a Linux/UNIX environment no
POSIX IDs are available for those objects.
</para>
<para>
The SID name space is organized in authorities which can be seen as
different domains. The authorities for the Well-Known SIDs are
<itemizedlist>
<listitem><para>Null Authority</para></listitem>
<listitem><para>World Authority</para></listitem>
<listitem><para>Local Authority</para></listitem>
<listitem><para>Creator Authority</para></listitem>
<listitem><para>Mandatory Label Authority</para></listitem>
<listitem><para>Authentication Authority</para></listitem>
<listitem><para>NT Authority</para></listitem>
<listitem><para>Built-in</para></listitem>
</itemizedlist>
The capitalized version of these names are used as domain names when
returning the fully qualified name of a Well-Known SID.
</para>
<para>
Since some utilities allow to modify SID based access control information
with the help of a name instead of using the SID directly SSSD supports to
look up the SID by the name as well. To avoid collisions only the fully
qualified names can be used to look up Well-Known SIDs. As a result the
domain names <quote>NULL AUTHORITY</quote>, <quote>WORLD AUTHORITY</quote>,
<quote>LOCAL AUTHORITY</quote>, <quote>CREATOR AUTHORITY</quote>,
<quote>MANDATORY LABEL AUTHORITY</quote>, <quote>AUTHENTICATION
AUTHORITY</quote>, <quote>NT AUTHORITY</quote> and <quote>BUILTIN</quote>
should not be used as domain names in <filename>sssd.conf</filename>.
</para>
</refsect2>
</refsect1>
|
