1SLAPD.ACCESS(5) File Formats Manual SLAPD.ACCESS(5)
2
6 slapd.access - access configuration for slapd, the stand-alone LDAP
7 daemon
8
10 /etc/openldap/slapd.conf
11
13 The slapd.conf(5) file contains configuration information for the
14 slapd(8) daemon. This configuration file is also used by the slurpd(8)
15 replication daemon and by the SLAPD tools slapacl(8), slapadd(8), sla‐
16 pauth(8), slapcat(8), slapdn(8), slapindex(8), and slaptest(8).
17 The slapd.conf file consists of a series of global configuration
19 options that apply to slapd as a whole (including all backends), fol‐
20 lowed by zero or more database backend definitions that contain infor‐
21 mation specific to a backend instance.
22 The general format of slapd.conf is as follows:
24 # comment - these options apply to every database
26 <global configuration options>
27 # first database definition & configuration options
28 database <backend 1 type>
29 <configuration options specific to backend 1>
30 # subsequent database definitions & configuration options
31 ...
32 Both the global configuration and each backend-specific section can
34 contain access information. Backend-specific access control directives
35 are used for those entries that belong to the backend, according to
36 their naming context. In case no access control directives are defined
37 for a backend or those which are defined are not applicable, the direc‐
38 tives from the global configuration section are then used.
39 If no access controls are present, the default policy allows anyone and
41 everyone to read anything but restricts updates to rootdn. (e.g.,
42 "access to * by * read"). The rootdn can always read and write EVERY‐
43 THING!
44 For entries not held in any backend (such as a root DSE), the direc‐
46 tives of the first backend (and any global directives) are used.
47 Arguments that should be replaced by actual text are shown in brackets
49 <>.
50
52 The structure of the access control directives is
53 access to <what> [ by <who> [ <access> ] [ <control> ] ]+
55 Grant access (specified by <access>) to a set of entries and/or
56 attributes (specified by <what>) by one or more requestors
57 (specified by <who>).
58
60 The field <what> specifies the entity the access control directive
61 applies to. It can have the forms
62 dn[.<dnstyle>]=<dnpattern>
64 filter=<ldapfilter>
65 attrs=<attrlist>[ val[/matchingRule][.<attrstyle>]=<attrval>]
66 with
68 <dnstyle>={{exact|base(object)}|regex
70 |one(level)|sub(tree)|children}
71 <attrlist>={<attr>|[{!|@}]<objectClass>}[,<attrlist>]
72 <attrstyle>={{exact|base(object)}|regex
73 |one(level)|sub(tree)|children}
74 The statement dn=<dnpattern> selects the entries based on their naming
76 context. The <dnpattern> is a string representation of the entry's DN.
77 The wildcard * stands for all the entries, and it is implied if no dn
78 form is given.
79 The <dnstyle> is optional; however, it is recommended to specify it to
81 avoid ambiguities. Base (synonym of baseObject), the default, or exact
82 (an alias of base) indicates the entry whose DN is equal to the <dnpat‐
83 tern>; one (synonym of onelevel) indicates all the entries immediately
84 below the <dnpattern>, sub (synonym of subtree) indicates all entries
85 in the subtree at the <dnpattern>, children indicates all the entries
86 below (subordinate to) the <dnpattern>.
87 If the <dnstyle> qualifier is regex, then <dnpattern> is a POSIX
89 (''extended'') regular expression pattern, as detailed in regex(7)
90 and/or re_format(7), matching a normalized string representation of the
91 entry's DN. The regex form of the pattern does not (yet) support
92 UTF-8.
93 The statement filter=<ldapfilter> selects the entries based on a valid
95 LDAP filter as described in RFC 2254. A filter of (objectClass=*) is
96 implied if no filter form is given.
97 The statement attrs=<attrlist> selects the attributes the access con‐
99 trol rule applies to. It is a comma-separated list of attribute types,
100 plus the special names entry, indicating access to the entry itself,
101 and children, indicating access to the entry's children. ObjectClass
102 names may also be specified in this list, which will affect all the
103 attributes that are required and/or allowed by that objectClass. Actu‐
104 ally, names in <attrlist> that are prefixed by @ are directly treated
105 as objectClass names. A name prefixed by ! is also treated as an
106 objectClass, but in this case the access rule affects the attributes
107 that are not required nor allowed by that objectClass. If no attrs
108 form is given, attrs=@extensibleObject is implied, i.e. all attributes
109 are addressed.
110 Using the form attrs=<attr> val[/matchingRule][.<attrstyle>]=<attrval>
112 specifies access to a particular value of a single attribute. In this
113 case, only a single attribute type may be given. The <attrstyle> exact
114 (the default) uses the attribute's equality matching rule to compare
115 the value, unless a different (and compatible) matching rule is speci‐
116 fied. If the <attrstyle> is regex, the provided value is used as a
117 POSIX (''extended'') regular expression pattern. If the attribute has
118 DN syntax, the <attrstyle> can be any of base, onelevel, subtree or
119 children, resulting in base, onelevel, subtree or children match,
120 respectively.
121 The dn, filter, and attrs statements are additive; they can be used in
123 sequence to select entities the access rule applies to based on naming
124 context, value and attribute type simultaneously.
125
127 The field <who> indicates whom the access rules apply to. Multiple
128 <who> statements can appear in an access control statement, indicating
129 the different access privileges to the same resource that apply to dif‐
130 ferent accessee. It can have the forms
131 *
133 anonymous
134 users
135 self[.<selfstyle>]
136 dn[.<dnstyle>[,<modifier>]]=<DN>
138 dnattr=<attrname>
139 realanonymous
141 realusers
142 realself[.<selfstyle>]
143 realdn[.<dnstyle>[,<modifier>]]=<DN>
145 realdnattr=<attrname>
146 group[/<objectclass>[/<attrname>]]
148 [.<groupstyle>]=<group>
149 peername[.<peernamestyle>]=<peername>
150 sockname[.<style>]=<sockname>
151 domain[.<domainstyle>[,<modifier>]]=<domain>
152 sockurl[.<style>]=<sockurl>
153 set[.<setstyle>]=<pattern>
154 ssf=<n>
156 transport_ssf=<n>
157 tls_ssf=<n>
158 sasl_ssf=<n>
159 aci[=<attrname>]
161 dynacl/name[/<options>][.<dynstyle>][=<pattern>]
162 with
164 <style>={exact|regex|expand}
166 <selfstyle>={level{<n>}}
167 <dnstyle>={{exact|base(object)}|regex
168 |one(level)|sub(tree)|children|level{<n>}}
169 <groupstyle>={exact|expand}
170 <peernamestyle>={<style>|ip|path}
171 <domainstyle>={exact|regex|sub(tree)}
172 <setstyle>={exact|regex}
173 <modifier>={expand}
174 They may be specified in combination.
176 The wildcard * refers to everybody.
178 The keywords prefixed by real act as their counterparts without prefix;
180 the checking respectively occurs with the authentication DN and the
181 authorization DN.
182 The keyword anonymous means access is granted to unauthenticated
184 clients; it is mostly used to limit access to authentication resources
185 (e.g. the userPassword attribute) to unauthenticated clients for
186 authentication purposes.
187 The keyword users means access is granted to authenticated clients.
189 The keyword self means access to an entry is allowed to the entry
191 itself (e.g. the entry being accessed and the requesting entry must be
192 the same). It allows the level{<n>} style, where <n> indicates what
193 ancestor of the DN is to be used in matches. A positive value indi‐
194 cates that the <n>-th ancestor of the user's DN is to be considered; a
195 negative value indicates that the <n>-th ancestor of the target is to
196 be considered. For example, a "by self.level{1} ..." clause would
197 match when the object "dc=example,dc=com" is accessed by
198 "cn=User,dc=example,dc=com". A "by self.level{-1} ..." clause would
199 match when the same user accesses the object "ou=Address
200 Book,cn=User,dc=example,dc=com".
201 The statement dn=<DN> means that access is granted to the matching DN.
203 The optional style qualifier dnstyle allows the same choices of the dn
204 form of the <what> field. In addition, the regex style can exploit
205 substring substitution of submatches in the <what> dn.regex clause by
206 using the form $<digit>, with digit ranging from 0 to 9 (where 0
207 matches the entire string), or the form ${<digit>+}, for submatches
208 higher than 9. Since the dollar character is used to indicate a sub‐
209 string replacement, the dollar character that is used to indicate match
210 up to the end of the string must be escaped by a second dollar charac‐
211 ter, e.g.
212 access to dn.regex="^(.+,)?uid=([^,]+),dc=[^,]+,dc=com$"
214 by dn.regex="^uid=$2,dc=[^,]+,dc=com$$" write
215 The style qualifier allows an optional modifier. At present, the only
217 type allowed is expand, which causes substring substitution of sub‐
218 matches to take place even if dnstyle is not regex. Note that the
219 regex dnstyle in the above example may be of use only if the <by>
220 clause needs to be a regex; otherwise, if the value of the second (from
221 the right) dc= portion of the DN in the above example were fixed, the
222 form
223 access to dn.regex="^(.+,)?uid=([^,]+),dc=example,dc=com$"
225 by dn.exact,expand="uid=$2,dc=example,dc=com" write
226 could be used; if it had to match the value in the <what> clause, the
228 form
229 access to dn.regex="^(.+,)?uid=([^,]+),dc=([^,]+),dc=com$"
231 by dn.exact,expand="uid=$2,dc=$3,dc=com" write
232 could be used.
234 Forms of the <what> clause other than regex may provide submatches as
236 well. The base(object), the sub(tree), the one(level), and the chil‐
237 dren forms provide $0 as the match of the entire string. The
238 sub(tree), the one(level), and the children forms also provide $1 as
239 the match of the rightmost part of the DN as defined in the <what>
240 clause. This may be useful, for instance, to provide access to all the
241 ancestors of a user by defining
242 access to dn.subtree="dc=com"
244 by dn.subtree,expand="$1" read
245 which means that only access to entries that appear in the DN of the
247 <by> clause is allowed.
248 The level{<n>} form is an extension and a generalization of the
250 onelevel form, which matches all DNs whose <n>-th ancestor is the pat‐
251 tern. So, level{1} is equivalent to onelevel, and level{0} is equiva‐
252 lent to base.
253 It is perfectly useless to give any access privileges to a DN that
255 exactly matches the rootdn of the database the ACLs apply to, because
256 it implicitly possesses write privileges for the entire tree of that
257 database. Actually, access control is bypassed for the rootdn, to
258 solve the intrinsic chicken-and-egg problem.
259 The statement dnattr=<attrname> means that access is granted to
261 requests whose DN is listed in the entry being accessed under the
262 <attrname> attribute.
263 The statement group=<group> means that access is granted to requests
265 whose DN is listed in the group entry whose DN is given by <group>.
266 The optional parameters <objectclass> and <attrname> define the object‐
267 Class and the member attributeType of the group entry. The defaults
268 are groupOfNames and member, respectively. The optional style quali‐
269 fier <style> can be expand, which means that <group> will be expanded
270 as a replacement string (but not as a regular expression) according to
271 regex(7) and/or re_format(7), and exact, which means that exact match
272 will be used. If the style of the DN portion of the <what> clause is
273 regex, the submatches are made available according to regex(7) and/or
274 re_format(7); other styles provide limited submatches as discussed
275 above about the DN form of the <by> clause.
276 For static groups, the specified attributeType must have Distinguished‐
278 Name or NameAndOptionalUID syntax. For dynamic groups the attributeType
279 must be a subtype of the labeledURI attributeType. Only LDAP URIs of
280 the form ldap:///<base>??<scope>?<filter> will be evaluated in a
281 dynamic group, by searching the local server only.
282 The statements peername=<peername>, sockname=<sockname>,
284 domain=<domain>, and sockurl=<sockurl> mean that the contacting host IP
285 (in the form IP=<ip>:<port>) or the contacting host named pipe file
286 name (in the form PATH=<path> if connecting through a named pipe) for
287 peername, the named pipe file name for sockname, the contacting host
288 name for domain, and the contacting URL for sockurl are compared
289 against pattern to determine access. The same style rules for pattern
290 match described for the group case apply, plus the regex style, which
291 implies submatch expand and regex match of the corresponding connection
292 parameters. The exact style of the <peername> clause (the default)
293 implies a case-exact match on the client's IP, including the IP= prefix
294 and the trailing :<port>, or the client's path, including the PATH=
295 prefix if connecting through a named pipe. The special ip style inter‐
296 prets the pattern as <peername>=<ip>[%<mask>][{<n>}], where <ip> and
297 <mask> are dotted digit representations of the IP and the mask, while
298 <n>, delimited by curly brackets, is an optional port. When checking
299 access privileges, the IP portion of the peername is extracted, elimi‐
300 nating the IP= prefix and the :<port> part, and it is compared against
301 the <ip> portion of the pattern after masking with <mask>. As an exam‐
302 ple, peername.ip=127.0.0.1 allows connections only from localhost,
303 peername.ip=192.168.1.0%255.255.255.0 allows connections from any IP in
304 the 192.168.1 class C domain, and peer‐
305 name.ip=192.168.1.16%255.255.255.240{9009} allows connections from any
306 IP in the 192.168.1.[16-31] range of the same domain, only if port 9009
307 is used. The special path style eliminates the PATH= prefix from the
308 peername when connecting through a named pipe, and performs an exact
309 match on the given pattern. The <domain> clause also allows the sub‐
310 tree style, which succeeds when a fully qualified name exactly matches
311 the domain pattern, or its trailing part, after a dot, exactly matches
312 the domain pattern. The expand style is allowed, implying an exact
313 match with submatch expansion; the use of expand as a style modifier is
314 considered more appropriate. As an example, domain.subtree=example.com
315 will match www.example.com, but will not match www.anotherexample.com.
316 The domain of the contacting host is determined by performing a DNS
317 reverse lookup. As this lookup can easily be spoofed, use of the
318 domain statement is strongly discouraged. By default, reverse lookups
319 are disabled. The optional domainstyle qualifier of the <domain>
320 clause allows a modifier option; the only value currently supported is
321 expand, which causes substring substitution of submatches to take place
322 even if the domainstyle is not regex, much like the analogous usage in
323 <dn> clause.
324 The statement set=<pattern> is undocumented yet.
326 The statement aci[=<attrname>] means that the access control is deter‐
328 mined by the values in the attrname of the entry itself. The optional
329 <attrname> indicates what attributeType holds the ACI information in
330 the entry. By default, the OpenLDAPaci operational attribute is used.
331 ACIs are experimental; they must be enabled at compile time.
332 The statement dynacl/<name>[/<options>][.<dynstyle>][=<pattern>] means
334 that access checking is delegated to the admin-defined method indicated
335 by <name>, which can be registered at run-time by means of the mod‐
336 uleload statement. The fields <options>, <dynstyle> and <pattern> are
337 optional, and are directly passed to the registered parsing routine.
338 Dynacl is experimental; it must be enabled at compile time. If dynacl
339 and ACIs are both enabled, ACIs are cast into the dynacl scheme, where
340 <name>=aci and, optionally, <patten>=<attrname>. However, the original
341 ACI syntax is preserved for backward compatibility.
342 The statements ssf=<n>, transport_ssf=<n>, tls_ssf=<n>, and
344 sasl_ssf=<n> set the minimum required Security Strength Factor (ssf)
345 needed to grant access. The value should be positive integer.
346
348 The field <access> ::= [[real]self]{<level>|<priv>} determines the
349 access level or the specific access privileges the who field will have.
350 Its component are defined as
351 <level> ::= none|disclose|auth|compare|search|read|write
353 <priv> ::= {=|+|-}{w|r|s|c|x|d|0}+
354 The modifier self allows special operations like having a certain
356 access level or privilege only in case the operation involves the name
357 of the user that's requesting the access. It implies the user that
358 requests access is authorized. The modifier realself refers to the
359 authenticated DN as opposed to the authorized DN of the self modifier.
360 An example is the selfwrite access to the member attribute of a group,
361 which allows one to add/delete its own DN from the member list of a
362 group, without affecting other members.
363 The level access model relies on an incremental interpretation of the
365 access privileges. The possible levels are none, disclose, auth, com‐
366 pare, search, read, and write. Each access level implies all the pre‐
367 ceding ones, thus write access will imply all accesses.
368 The none access level disallows all access including disclosure on
370 error.
371 The disclose access level allows disclosure of information on error.
373 The auth access level means that one is allowed access to an attribute
375 to perform authentication/authorization operations (e.g. bind) with no
376 other access. This is useful to grant unauthenticated clients the
377 least possible access level to critical resources, like passwords.
378 The priv access model relies on the explicit setting of access privi‐
380 leges for each clause. The = sign resets previously defined accesses;
381 as a consequence, the final access privileges will be only those
382 defined by the clause. The + and - signs add/remove access privileges
383 to the existing ones. The privileges are w for write, r for read, s
384 for search, c for compare, x for authentication, and d for disclose.
385 More than one of the above privileges can be added in one statement. 0
386 indicates no privileges and is used only by itself (e.g., +0). If no
387 access is given, it defaults to +0.
388
390 The optional field <control> controls the flow of access rule applica‐
391 tion. It can have the forms
392 stop
394 continue
395 break
396 where stop, the default, means access checking stops in case of match.
398 The other two forms are used to keep on processing access clauses. In
399 detail, the continue form allows for other <who> clauses in the same
400 <access> clause to be considered, so that they may result in incremen‐
401 tally altering the privileges, while the break form allows for other
402 <access> clauses that match the same target to be processed. Consider
403 the (silly) example
404 access to dn.subtree="dc=example,dc=com" attrs=cn
406 by * =cs break
407 access to dn.subtree="ou=People,dc=example,dc=com"
409 by * +r
410 which allows search and compare privileges to everybody under the
412 "dc=example,dc=com" tree, with the second rule allowing also read in
413 the "ou=People" subtree, or the (even more silly) example
414 access to dn.subtree="dc=example,dc=com" attrs=cn
416 by * =cs continue
417 by users +r
418 which grants everybody search and compare privileges, and adds read
420 privileges to authenticated clients.
421 One useful application is to easily grant write privileges to an updat‐
423 edn that is different from the rootdn. In this case, since the updat‐
424 edn needs write access to (almost) all data, one can use
425 access to *
427 by dn.exact="cn=The Update DN,dc=example,dc=com" write
428 by * break
429 as the first access rule. As a consequence, unless the operation is
431 performed with the updatedn identity, control is passed straight to the
432 subsequent rules.
433
436 Operations require different privileges on different portions of
437 entries. The following summary applies to primary database backends
438 such as the BDB and HDB backends. Requirements for other backends may
439 (and often do) differ.
440 The add operation requires write (=w) privileges on the pseudo-
443 attribute entry of the entry being added, and write (=w) privileges on
444 the pseudo-attribute children of the entry's parent. When adding the
445 suffix entry of a database, write access to children of the empty DN
446 ("") is required.
447 The bind operation, when credentials are stored in the directory,
450 requires auth (=x) privileges on the attribute the credentials are
451 stored in (usually userPassword).
452 The compare operation requires compare (=c) privileges on the attribute
455 that is being compared.
456 The delete operation requires write (=w) privileges on the pseudo-
459 attribute entry of the entry being deleted, and write (=w) privileges
460 on the children pseudo-attribute of the entry's parent.
461 The modify operation requires write (=w) privileges on the attributes
464 being modified.
465 The modrdn operation requires write (=w) privileges on the pseudo-
468 attribute entry of the entry whose relative DN is being modified, write
469 (=w) privileges on the pseudo-attribute children of the old and new
470 entry's parents, and write (=w) privileges on the attributes that are
471 present in the new relative DN. Write (=w) privileges are also
472 required on the attributes that are present in the old relative DN if
473 deleteoldrdn is set to 1.
474 The search operation, requires search (=s) privileges on the entry
477 pseudo-attribute of the searchBase (NOTE: this was introduced with
478 2.3). Then, for each entry, it requires search (=s) privileges on the
479 attributes that are defined in the filter. The resulting entries are
480 finally tested for read (=r) privileges on the pseudo-attribute entry
481 (for read access to the entry itself) and for read (=r) access on each
482 value of each attribute that is requested. Also, for each referral
483 object used in generating continuation references, the operation
484 requires read (=r) access on the pseudo-attribute entry (for read
485 access to the referral object itself), as well as read (=r) access to
486 the attribute holding the referral information (generally the ref
487 attribute).
488 Some internal operations and some controls require specific access
491 privileges. The authzID mapping and the proxyAuthz control require
492 auth (=x) privileges on all the attributes that are present in the
493 search filter of the URI regexp maps (the right-hand side of the authz-
494 regexp directives). Auth (=x) privileges are also required on the
495 authzTo attribute of the authorizing identity and/or on the authzFrom
496 attribute of the authorized identity.
497 Access control to search entries is checked by the frontend, so it is
500 fully honored by all backends; for all other operations and for the
501 discovery phase of the search operation, full ACL semantics is only
502 supported by the primary backends, i.e. back-bdb(5), and back-hdb(5).
503 Some other backend, like back-sql(5), may fully support them; others
505 may only support a portion of the described semantics, or even differ
506 in some aspects. The relevant details are described in the backend-
507 specific man pages.
508
511 It is strongly recommended to explicitly use the most appropriate
512 <dnstyle> in <what> and <who> clauses, to avoid possible incorrect
513 specifications of the access rules as well as for performance (avoid
514 unnecessary regex matching when an exact match suffices) reasons.
515 An administrator might create a rule of the form:
517 access to dn.regex="dc=example,dc=com"
519 by ...
520 expecting it to match all entries in the subtree "dc=example,dc=com".
522 However, this rule actually matches any DN which contains anywhere the
523 substring "dc=example,dc=com". That is, the rule matches both
524 "uid=joe,dc=example,dc=com" and "dc=example,dc=com,uid=joe".
525 To match the desired subtree, the rule would be more precisely written:
527 access to dn.regex="^(.+,)?dc=example,dc=com$"
529 by ...
530 For performance reasons, it would be better to use the subtree style.
532 access to dn.subtree="dc=example,dc=com"
534 by ...
535 When writing submatch rules, it may be convenient to avoid unnecessary
537 regex <dnstyle> use; for instance, to allow access to the subtree of
538 the user that matches the <what> clause, one could use
539 access to dn.regex="^(.+,)?uid=([^,]+),dc=example,dc=com$"
541 by dn.regex="^uid=$2,dc=example,dc=com$$" write
542 by ...
543 However, since all that is required in the <by> clause is substring
545 expansion, a more efficient solution is
546 access to dn.regex="^(.+,)?uid=([^,]+),dc=example,dc=com$"
548 by dn.exact,expand="uid=$2,dc=example,dc=com" write
549 by ...
550 In fact, while a <dnstyle> of regex implies substring expansion, exact,
552 as well as all the other DN specific <dnstyle> values, does not, so it
553 must be explicitly requested.
554
556 /etc/openldap/slapd.conf
557 default slapd configuration file
558
560 slapd(8), slapd-*[22m(5), slapacl(8), regex(7), re_format(7)
561 "OpenLDAP Administrator's Guide" (http://www.OpenLDAP.org/doc/admin/)
563
565 OpenLDAP is developed and maintained by The OpenLDAP Project
566 (http://www.openldap.org/). OpenLDAP is derived from University of
567 Michigan LDAP 3.3 Release.
568OpenLDAP 2.3.34 2007/2/16 SLAPD.ACCESS(5)