OpenID Connect Federation Protocol
The CAS server supports the OpenID Federation protocol.
It can act as only one of the following roles (mutually exclusive):
- Trust Anchor (TA)
- Intermediate
- OpenID Provider (OP)
Support is enabled by including the following dependency in the WAR overlay:
1
2
3
4
5
<dependency>
<groupId>org.apereo.cas</groupId>
<artifactId>cas-server-support-oidc-federation</artifactId>
<version>${cas.version}</version>
</dependency>
1
implementation "org.apereo.cas:cas-server-support-oidc-federation:${project.'cas.version'}"
1
2
3
4
5
6
7
8
9
dependencyManagement {
imports {
mavenBom "org.apereo.cas:cas-server-support-bom:${project.'cas.version'}"
}
}
dependencies {
implementation "org.apereo.cas:cas-server-support-oidc-federation"
}
1
2
3
4
5
6
7
8
9
10
dependencies {
/*
The following platform references are included automatically and are listed for reference only.
implementation enforcedPlatform("org.apereo.cas:cas-server-support-bom:${project.'cas.version'}")
implementation platform(org.springframework.boot.gradle.plugin.SpringBootPlugin.BOM_COORDINATES)
*/
implementation "org.apereo.cas:cas-server-support-oidc-federation"
}
Configuration
The following settings and properties are available from the CAS configuration catalog:
cas.authn.oidc.federation.authority-hintsThe List of federation authorities where this entity is enrolled, i.e. its immediate superiors.
The List of federation authorities where this entity is enrolled, i.e. its immediate superiors. This is used to generate the federation metadata. When no trust anchors are defined, the entries are also considered trust anchors when validating incoming federation requests from relying parties.
cas.authn.oidc.federation.contactsList of contacts to include in the federation metadata.
List of contacts to include in the federation metadata.
cas.authn.oidc.federation.jwks-filePath to the JWKS file resource used to handle signing of federation metadata.
file:/etc/cas/config/federation-keystore.jwksPath to the JWKS file resource used to handle signing of federation metadata. Note that if the keystore file does not exist at the specified path, one will be generated for you, with a current and a future key. Keys are rotated and retired in this file, so it must be writable when rotation or revocation is used.
cas.authn.oidc.federation.organizationOrganization name to include in the federation metadata.
Apereo CASOrganization name to include in the federation metadata.
cas.authn.oidc.federation.roleThe role of this server.
-
TRUST_ANCHOR: The root of trust entity issuing statements about other entities. -
INTERMEDIATE: Intermediate entity between the trust anchor and the leaf entities. -
OPENID_PROVIDER: Leaf entity playing the OpenID Connect provider protocol role.
cas.authn.oidc.federation.entity-statement-expirationSets the time period during which the entity statement is not expired.
P90DSets the time period during which the entity statement is not expired.
cas.authn.oidc.federation.issued-trust-marks
cas.authn.oidc.federation.issued-trust-marks[0].delegation
cas.authn.oidc.federation.issued-trust-marks[0].expiration
P30Dcas.authn.oidc.federation.issued-trust-marks[0].logo-uri
cas.authn.oidc.federation.issued-trust-marks[0].ref
cas.authn.oidc.federation.issued-trust-marks[0].subjects
cas.authn.oidc.federation.issued-trust-marks[0].trust-mark-type
cas.authn.oidc.federation.max-authority-hints
5cas.authn.oidc.federation.max-concurrent-resolutions
10cas.authn.oidc.federation.max-path-length
3cas.authn.oidc.federation.max-requests-per-resolution
50cas.authn.oidc.federation.required-trust-marks
cas.authn.oidc.federation.resolution-timeout
PT10Scas.authn.oidc.federation.resolve-endpoint-authentication
nonecas.authn.oidc.federation.revocation.schedule.cron-expression
cas.authn.oidc.federation.revocation.schedule.cron-time-zone
cas.authn.oidc.federation.revocation.schedule.enabled
truecas.authn.oidc.federation.revocation.schedule.enabled-on-host
.*cas.authn.oidc.federation.revocation.schedule.repeat-interval
PT2Mcas.authn.oidc.federation.revocation.schedule.start-delay
PT15Scas.authn.oidc.federation.rotation.schedule.cron-expression
cas.authn.oidc.federation.rotation.schedule.cron-time-zone
cas.authn.oidc.federation.rotation.schedule.enabled
truecas.authn.oidc.federation.rotation.schedule.enabled-on-host
.*cas.authn.oidc.federation.rotation.schedule.repeat-interval
PT2Mcas.authn.oidc.federation.rotation.schedule.start-delay
PT15Scas.authn.oidc.federation.subordinate-directoryFor a TA or intermediate, the file directory where the subordinates are stored.
For a TA or intermediate, the file directory where the subordinates are stored.
cas.authn.oidc.federation.trust-anchors
cas.authn.oidc.federation.trust-anchors[0].entity-id
cas.authn.oidc.federation.trust-anchors[0].jwks
cas.authn.oidc.federation.trust-mark-endpoint-authentication
nonecas.authn.oidc.federation.trust-mark-issuers
cas.authn.oidc.federation.trust-mark-issuers[0].issuers
cas.authn.oidc.federation.trust-mark-issuers[0].trust-mark-type
cas.authn.oidc.federation.trust-mark-owners
cas.authn.oidc.federation.trust-mark-owners[0].jwks
cas.authn.oidc.federation.trust-mark-owners[0].owner
cas.authn.oidc.federation.trust-mark-owners[0].trust-mark-type
cas.authn.oidc.federation.trust-marks
cas.authn.oidc.federation.unresolved-entity-retention
PT5MRequired settings may be needed to activate or affect the feature; review them even when they have a default. Optional settings only need to be set to change a default or to turn on the behavior they control. Third party settings belong to libraries such as Spring Boot that CAS builds on; their own documentation may have more detail.
Notes on configuration
Configuration Metadata
The collection of configuration properties listed in this section are automatically generated from the CAS source and components that contain the actual field definitions, types, descriptions, modules, etc. This metadata may not always be 100% accurate, or could be lacking details and sufficient explanations.
Be Selective
This section is meant as a guide only. Do NOT copy/paste the entire collection of settings into your CAS configuration; rather pick only the properties that you need. Do NOT enable settings unless you are certain of their purpose and do NOT copy settings into your configuration only to keep them as reference. All these ideas lead to upgrade headaches, maintenance nightmares and premature aging.
YAGNI
Note that for nearly ALL use cases, declaring and configuring properties listed here is sufficient. You should NOT have to explicitly massage a CAS XML/Java/etc configuration file to design an authentication handler, create attribute release policies, etc. CAS at runtime will auto-configure all required changes for you. If you are unsure about the meaning of a given CAS setting, do NOT turn it on without hesitation. Review the codebase or better yet, ask questions to clarify the intended behavior.
Naming Convention
Property names can be specified in very relaxed terms. For instance cas.someProperty, cas.some-property, cas.some_property are all valid names. While all
forms are accepted by CAS, there are certain components (in CAS and other frameworks used) whose activation at runtime is conditional on a property value, where
this property is required to have been specified in CAS configuration using kebab case. This is both true for properties that are owned by CAS as well as those
that might be presented to the system via an external library or framework such as Spring Boot, etc.
When possible, properties should be stored in lower-case kebab format, such as cas.property-name=value.
The only possible exception to this rule is when naming actuator endpoints; The name of the
actuator endpoints (i.e. ssoSessions) MUST remain in camelCase mode.
Settings and properties that are controlled by the CAS platform directly always begin with the prefix cas. All other settings are controlled and provided
to CAS via other underlying frameworks and may have their own schemas and syntax. BE CAREFUL with
the distinction. Unrecognized properties are rejected by CAS and/or frameworks upon which CAS depends. This means if you somehow misspell a property definition
or fail to adhere to the dot-notation syntax and such, your setting is entirely refused by CAS and likely the feature it controls will never be activated in the
way you intend.
Validation
Configuration properties are automatically validated on CAS startup to report issues with configuration binding, especially if defined CAS settings cannot be recognized or validated by the configuration schema. Additional validation processes are also handled via Configuration Metadata and property migrations applied automatically on startup by Spring Boot and family.
Indexed Settings
CAS settings able to accept multiple values are typically documented with an index, such as cas.some.setting[0]=value. The index [0] is meant to be
incremented by the adopter to allow for distinct multiple configuration blocks.
Subordinates
In OpenID Connect Federation, a subordinate is an entity that is “below” another federation entity in the trust hierarchy. Simply put, a subordinate is an entity whose federation metadata and trust relationship are vouched for by a higher-level entity in the federation.
More concretely, a subordinate is an entity for which a superior entity, such as a Trust Anchor or Intermediate Entity, can issue a Subordinate Statement. That statement is a signed Entity Statement saying, in effect:
“I, the issuer, make authoritative claims about this subject entity.”
The iss is the superior, and the sub is the subordinate.
1
2
3
4
Trust Anchor
└── Intermediate Entity
└── OpenID Provider
└── Relying Party
For a trust anchor or intermediate, the supported subordinates must be defined
as JSON files with the OidcFederationSubordinate format:
Relying Party Subordinate
Here is an example of an RP subordinate definition:
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
{
"entityId" : "https://rp.example.org",
"metadata": {
"openid_relying_party": {
"redirect_uris": [
"https://rp.example.org/callback"
],
"application_type": "web",
"response_types": [
"code"
],
"grant_types": [
"authorization_code"
],
"scope": "openid email profile",
"token_endpoint_auth_method": "private_key_jwt",
"token_endpoint_auth_signing_alg": "RS256",
"request_object_signing_alg": "RS256",
"jwks": {
"keys": [
{
"kty": "RSA",
"e": "AQAB",
"use": "sig",
"kid": "keyid",
"n": "2moVQ...2aq7Q"
}
]
},
"client_registration_types": [
"explicit", "automatic"
],
"client_name": "RP test"
}
},
"federationKeys": [
{
"kty": "EC",
"use": "sig",
"crv": "P-256",
"kid": "TJns",
"x": "iLPMKMGqb7HeCnov5xBwEnq2zUbX5gNbyoVn1EHpvSI",
"y": "1kcVTdS1_Y2_NnL0E9LpWhPWsoCNOFQR2Dd7VAJBYKY"
}
]
}
OpenId Provider Subordinate
Here is an example of an OP subordinate definition, with a shortened openid_provider metadata:
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
{
"entityId" : "https://op.example.org",
"metadata": {
"openid_provider": {
"issuer": "https://op.example.org",
"authorization_endpoint": "https://op.example.org/authorize",
"token_endpoint": "https://op.example.org/token",
"pushed_authorization_request_endpoint": "https://op.example.org/par",
"jwks_uri": "https://op.example.org/jwks.json",
"client_registration_types_supported": [ "automatic" ],
"scopes_supported": [ "openid", "profile", "email" ],
"response_types_supported": [ "code" ],
"grant_types_supported": [ "authorization_code", "refresh_token" ],
"token_endpoint_auth_methods_supported": [ "private_key_jwt" ],
"request_object_signing_alg_values_supported": [ "RS256", "ES256" ]
}
},
"federationKeys": [
{
"kty": "RSA",
"e": "AQAB",
"use": "sig",
"kid": "PRRz",
"n": "vPZc7Y..."
}
]
}
The metadata and federationKeys properties come both from the
data in the .well-known/openid-federation endpoint. The federationKeys section is mandatory in the federation
entity definitions, while metadata is optional. Metadata in a subordinate statement takes precedence over the
metadata the entity publishes itself, so an empty object for an entity type, such as "openid_relying_party": {},
leaves the entity’s own metadata in effect.
Metadata Policies
A subordinate definition may also carry a metadata policy, which CAS publishes in its subordinate statement as
OpenID Federation describes, to restrict or adjust
the metadata of the subordinate and of every entity below it. Policies use the standard operators value, add,
default, one_of, subset_of, superset_of and essential; operators beyond those are ignored by entities that
do not understand them, unless they are listed as critical:
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
{
"entityId" : "https://rp.example.org",
"metadataPolicy": {
"openid_relying_party": {
"scope": {
"subset_of": [ "openid", "email", "profile" ]
},
"contacts": {
"add": [ "ops@example.org" ]
},
"token_endpoint_auth_method": {
"one_of": [ "private_key_jwt", "tls_client_auth" ],
"essential": true
}
}
},
"metadataPolicyCrit": [],
"federationKeys": [
{
"kty": "RSA",
"e": "AQAB",
"use": "sig",
"kid": "PRRz",
"n": "vPZc7Y..."
}
]
}
A metadata policy that CAS cannot read, such as one that combines operators in a way the specification forbids, is not published. Subordinates are listed for the entity types of their metadata as well as of their metadata policy.
A subordinate definition may also carry constraints, published as they are in its subordinate statement, as OpenID
Federation describes, to bound the trust chains that
go through the subordinate:
1
2
3
4
5
6
7
8
9
10
11
{
"entityId" : "https://intermediate.example.org",
"constraints": {
"max_path_length": 1,
"naming_constraints": {
"permitted": [ ".example.org" ]
},
"allowed_entity_types": [ "openid_relying_party" ]
},
"federationKeys": [ ]
}
The fetch, list and trust mark status endpoints of a subordinate are left out of the federation_entity metadata of its
subordinate statement, since they may only be published in the entity configuration of the subordinate.
Signed entity statements are returned from the /oidc/fetch endpoint, depending on the requested subordinate (sub
request parameter). A subordinate statement carries the keys of the subordinate only. An unknown subordinate is answered
with 404 and not_found, a missing sub or the entity identifier of CAS itself with 400 and invalid_request, and
a subordinate defined without keys, with a metadata policy CAS cannot read, or with constraints that are not an object,
with 500 and server_error; entities that issue no subordinate statements answer 404.
The /oidc/list endpoint lists the subordinates, and can filter them by entity type, with one or more entity_type
request parameters, and to intermediates only, with intermediate=true. A subordinate has the entity types its
definition lists in entityTypes, or, when it lists none, the entity types of its metadata; a metadata policy for an
entity type does not give the subordinate that type, so a definition that only carries a policy should list its types.
Filtering by trust marks is not supported and is answered with unsupported_parameter.
Resolve Endpoint
As a trust anchor or an intermediate, CAS resolves entities of the federation for others at its /oidc/resolve
endpoint, advertised as federation_resolve_endpoint, as described by OpenID
Federation. A GET request names the entity with
sub, one or more trust anchors with trust_anchor and, optionally, the entity types whose metadata is wanted with one
or more entity_type; all entity types are returned otherwise.
Requests are not authenticated by default. Client authentication can be made optional or required using CAS
properties, as described by OpenID Federation: an
authenticated request is a POST with the same parameters and a private_key_jwt client assertion, signed with a
federation key of a subordinate of CAS as its subordinate definition holds it, whose iss and sub are the
subordinate, whose aud is only the entity identifier of CAS, and which carries jti and exp; each client assertion
is accepted once. The resolve response is then addressed to the subordinate with aud. The entity configuration
advertises the accepted methods as federation_resolve_endpoint_auth_methods. A GET when authentication is required,
or a request that does not authenticate, is answered with 401 and invalid_client; a POST when authentication is
not supported with 400 and invalid_request.
- A trust anchor resolves entities to itself. An intermediate resolves them to its trust anchors, or to its authority hints when no trust anchor is defined. Requested trust anchors that CAS does not trust are ignored.
- CAS resolves the trust chains of the entity with the same bounds on authority hints and path length as for automatic registration, and builds its own statements itself rather than requesting them from its own endpoints. The shortest trust chain that satisfies the constraints of its statements and is not signed with a key its trust anchor has revoked is used, ending with the entity configuration of the trust anchor.
- The metadata of each entity type is resolved with the metadata policies of the trust chain. Entity types the constraints of the chain do not allow are left out.
- Only the trust marks of the entity that are valid, issued by issuers the trust anchor accepts for their type and active, are included. The response expires with the trust chain, or with the first of those trust marks to expire. A trust mark that cannot be checked for now, such as when its status endpoint cannot be reached, is left out, and the response is then not remembered.
- The response is a JWT of type
resolve-response+jwt, signed with the current federation key. - Resolved trust chains, and entities that could not be resolved, are remembered per entity and trust anchors for the period that applies to entity identifiers that could not be resolved, or until the chain or a trust mark expires, so that repeated requests do not resolve them again. That period can be tuned using CAS properties.
- Errors are answered as JSON: a missing or malformed parameter with
400andinvalid_request; trust anchors CAS does not trust with404andinvalid_trust_anchor; an entity without a usable trust chain to them with400andinvalid_trust_chain; metadata the policies of the chain cannot apply to with400andinvalid_metadata; a resolution that cannot start because too many are running, or that cannot tell about a trust anchor that cannot be reached, answers with a server error or whose historical keys cannot be read, with503andtemporarily_unavailable, which is not remembered. An OpenID Provider answers404andnot_found.
Entity Configuration
The entity identifier of CAS is its issuer, without a trailing slash, in every role.
CAS publishes its entity configuration at .well-known/openid-federation under its issuer, signed with its federation
keys, whatever its role. As an OpenID Provider, it describes itself as openid_provider with its discovery metadata,
which advertises automatic as the only client registration type; plain OpenID Connect deployments without the
federation role do not advertise client registration types. When CAS also
issues verifiable credentials, the entity configuration describes it
as openid_credential_issuer, with the same metadata as its credential issuer metadata endpoint, and as
oauth_authorization_server, which the credential issuer names as its authorization server, so that wallets of
federation-based ecosystems find and trust the credential issuer through its trust chain
(OpenID Federation for Wallet Architectures). When CAS
presents itself to wallets as a verifier with the openid_federation client identifier prefix, the entity
configuration also describes it as openid_credential_verifier, with its keys and response URI, as described in
verifiable presentations. The credential
issuer and authorization server entity types require the credential issuer of CAS to be the entity identifier;
otherwise they are left out, with a warning.
Trust marks that trust mark issuers have issued to CAS are published in the entity configuration as trust_marks. Each
trust mark is a signed JWT, given as a file, a URL or the JWT itself; files and URLs are read again each time the entity
configuration is produced, so a renewed trust mark is published without a restart. The trust mark type is taken from
the trust mark, and a trust mark that is about another entity, names no type or has expired is not published and is
reported in the logs. Trust marks are not verified by CAS, which leaves that to the parties that resolve it, and can be
defined using CAS properties.
As a trust anchor, CAS publishes in its entity configuration the trust mark types of the federation and the entities it
accepts as their issuers, as trust_mark_issuers. A trust mark type listed without issuers may be issued by anyone.
Both can be defined using CAS properties. A trust anchor also publishes, as trust_mark_owners, the owners of trust
mark types whose issuance they delegate to other issuers, each with the keys their delegations are verified with; only
the public members of those keys are published.
Trust Mark Issuer
In any role, CAS can issue trust marks, as described by OpenID Federation. Each trust mark type is defined with the entities it is issued to, how long its trust marks are valid, and optionally a logo, a reference to human-readable information and a delegation from the owner of the type, all of which can be defined using CAS properties. Nothing is stored: a trust mark is signed with the current federation key when it is requested. When CAS issues trust marks, its entity configuration advertises three endpoints, which do not require authentication by default:
- The trust mark endpoint answers a request for a trust mark type and an entity with a trust mark of that type about
that entity, typed
trust-mark+jwt, or with404when the entity does not hold that type. - The trust mark status endpoint answers the status of a trust mark it was sent, in a signed status response:
active;expiredonce the trust mark has expired;revokedwhen its subject no longer holds the type; orinvalidwhen it is not typed as a trust mark or does not verify with the federation keys CAS publishes. A trust mark issued by another entity, or of a type CAS does not issue, is answered with404. - The trust marked entities listing endpoint lists the entities that hold a trust mark type, optionally only the one asked for.
Requests to the trust mark endpoint can also be authenticated, as
OpenID Federation allows, so that entities only
obtain their own trust marks. Client authentication at that endpoint can be left off, which is the default, allowed, or
required, using CAS properties, and the entity configuration advertises the choice along with the signing algorithms
CAS accepts. An authenticated request is a POST that authenticates with private_key_jwt: the client assertion is
signed with a federation key of the requester, as its subordinate definition holds it; its issuer and
subject are the requester; its only audience is the entity identifier of CAS; and it is accepted once, until it
expires. Only subordinates of CAS as a trust anchor or intermediate can therefore authenticate, and they may only ask
for their own trust marks. A request that fails to authenticate, or asks for the trust mark of another entity, is
answered with 401 and invalid_client, as is an unauthenticated request when client authentication is required.
Removing an entity from a trust mark type revokes the trust marks of that type it holds. A trust mark signed with a
federation key that has since been retired stays active for the rest of its life, since the historical keys keep it
verifiable, as OpenID Federation intends; it becomes
invalid once that key is revoked.
Key Rollover
CAS signs its entity configuration and the statements it issues with its federation keys, which are kept in their own keystore and go through the same lifecycle as the OpenID Connect keys, as OpenID Federation describes for rolling keys over:
- A new keystore starts with a current key, which signs, and a future key, which is only published. A keystore that holds a single key without a lifecycle state keeps working; its key is current. A keystore that holds no keys is given new ones, and one that cannot be read stops CAS from starting rather than being replaced.
- The entity configuration publishes the current, future and previous keys, so a key is known to the federation before it starts signing and after it stopped.
- Rotation makes the future key current, the current key previous, and publishes a new future key. When the keystore has no future key yet, rotation only publishes one, and the next rotation makes it current.
- Revocation retires the previous keys: they leave the entity configuration and are kept, without their private keys, as historical keys, with the time they were retired as their expiration.
- Every role publishes its historical keys at the federation historical keys endpoint, advertised in its entity configuration, as a JWK set signed with the current key.
- A key can also be revoked, for instance when its private key is believed compromised. A key that is still published
leaves the entity configuration at once and is published as a historical key that expires at the time of the
revocation and is marked
revoked, with that time and the reason, such ascompromised,supersededor a reason defined by the federation. A key that was already retired keeps its expiration and is marked revoked as well. When the revoked key was signing, the future key starts signing, or a new key when there is none, and a new future key is published. A key that signs without having been published before is unknown to superiors until they update their subordinate statements about CAS.
Rotation and revocation can run on a schedule, which can be defined using CAS properties, or be triggered on demand, which is also how a key is revoked. A scheduled run only changes keys once they are old enough, at half the time between two runs of its schedule: a scheduled rotation makes the future key current only once that key has been published that long, and a scheduled revocation retires a previous key only once it has been previous that long. A schedule that runs again soon after another run leaves the keys as they are: after a restart, or on another node that shares the keystore file and reads it after the change. Nodes whose schedules run at the very same moment may each change the keys, so in a cluster the schedules are best enabled on a single node. On demand, rotation and revocation apply at once:
oidcFederationJwks
- Returns
ResponseEntity- Produces
application/json- Java method
revoke()- Defined in
OidcFederationJwksEndpoint
| Name | In | Type | Description |
|---|---|---|---|
kidrequired
|
path | – | The key identifier |
reason |
query | – | The reason: compromised, superseded, unspecified or a reason of the federation |
- Returns
ResponseEntity- Produces
application/json- Java method
revokeKey(String,String)- Defined in
OidcFederationJwksEndpoint
- Returns
ResponseEntity- Produces
application/json- Java method
rotate()- Defined in
OidcFederationJwksEndpoint
Include the module that provides this endpoint in the WAR overlay:
1
2
3
4
5
<dependency>
<groupId>org.apereo.cas</groupId>
<artifactId>cas-server-support-oidc-federation</artifactId>
<version>${cas.version}</version>
</dependency>
1
implementation "org.apereo.cas:cas-server-support-oidc-federation:${project.'cas.version'}"
1
2
3
4
5
6
7
8
9
dependencyManagement {
imports {
mavenBom "org.apereo.cas:cas-server-support-bom:${project.'cas.version'}"
}
}
dependencies {
implementation "org.apereo.cas:cas-server-support-oidc-federation"
}
1
2
3
4
5
6
7
8
9
10
dependencies {
/*
The following platform references are included automatically and are listed for reference only.
implementation enforcedPlatform("org.apereo.cas:cas-server-support-bom:${project.'cas.version'}")
implementation platform(org.springframework.boot.gradle.plugin.SpringBootPlugin.BOM_COORDINATES)
*/
implementation "org.apereo.cas:cas-server-support-oidc-federation"
}
Turn the endpoint on and expose it over the web.
One entry covers every operation. By default only info, health and status are exposed.
1
2
management.endpoint.oidcFederationJwks.access=UNRESTRICTED
management.endpoints.web.exposure.include=oidcFederationJwks
1
2
3
4
5
6
7
8
management:
endpoint:
oidcFederationJwks:
access: "UNRESTRICTED"
endpoints:
web:
exposure:
include: "oidcFederationJwks"
1
2
MANAGEMENT_ENDPOINT_OIDCFEDERATIONJWKS_ACCESS=UNRESTRICTED
MANAGEMENT_ENDPOINTS_WEB_EXPOSURE_INCLUDE=oidcFederationJwks
Endpoints may be mapped to other paths. For example, to serve health at healthcheck:
1
management.endpoints.web.path-mapping.health=healthcheck
Once exposed, every call goes through Spring Security,
and CAS decides access per endpoint. Without a rule, the special defaults endpoint applies, which denies access.
1
cas.monitor.endpoints.endpoint.oidcFederationJwks.access=AUTHENTICATED
1
2
3
4
5
6
cas:
monitor:
endpoints:
endpoint:
oidcFederationJwks:
access: "AUTHENTICATED"
1
CAS_MONITOR_ENDPOINTS_ENDPOINT_OIDCFEDERATIONJWKS_ACCESS=AUTHENTICATED
1
2
cas.monitor.endpoints.endpoint.oidcFederationJwks.access=ROLE
cas.monitor.endpoints.endpoint.oidcFederationJwks.required-roles=ADMIN
1
2
3
4
5
6
7
cas:
monitor:
endpoints:
endpoint:
oidcFederationJwks:
access: "ROLE"
required-roles: "ADMIN"
1
2
CAS_MONITOR_ENDPOINTS_ENDPOINT_OIDCFEDERATIONJWKS_ACCESS=ROLE
CAS_MONITOR_ENDPOINTS_ENDPOINT_OIDCFEDERATIONJWKS_REQUIREDROLES=ADMIN
1
2
cas.monitor.endpoints.endpoint.oidcFederationJwks.access=IP_ADDRESS
cas.monitor.endpoints.endpoint.oidcFederationJwks.required-ip-addresses=127.0.0.1
1
2
3
4
5
6
7
cas:
monitor:
endpoints:
endpoint:
oidcFederationJwks:
access: "IP_ADDRESS"
required-ip-addresses: "127.0.0.1"
1
2
CAS_MONITOR_ENDPOINTS_ENDPOINT_OIDCFEDERATIONJWKS_ACCESS=IP_ADDRESS
CAS_MONITOR_ENDPOINTS_ENDPOINT_OIDCFEDERATIONJWKS_REQUIREDIPADDRESSES=127.0.0.1
1
cas.monitor.endpoints.endpoint.oidcFederationJwks.access=PERMIT
1
2
3
4
5
6
cas:
monitor:
endpoints:
endpoint:
oidcFederationJwks:
access: "PERMIT"
1
CAS_MONITOR_ENDPOINTS_ENDPOINT_OIDCFEDERATIONJWKS_ACCESS=PERMIT
A single user defined in CAS settings. If the name and password are left undefined, CAS generates them at startup.
spring.security.user.nameDefault user name.
userDefault user name.
spring.security.user.passwordPassword for the default user name.
Password for the default user name.
spring.security.user.rolesGranted roles for the default user name.
Granted roles for the default user name.
cas.monitor.endpoints.jaas.login-configJAAS login resource file.
JAAS login resource file.
cas.monitor.endpoints.jaas.login-context-nameThe login context name should coincide with a given index in the login config specified.
The login context name should coincide with a given index in the login config specified. This name is used as the index to the configuration specified in the login config property.
JAASTest { org.springframework.security.authentication.jaas.TestLoginModule required; }; In the above example, JAASTest should be set as the context name.cas.monitor.endpoints.jaas.refresh-configuration-on-startupIf set, a call to Configuration#refresh() will be made by #configureJaas(Resource) method.
trueIf set, a call to Configuration#refresh() will be made by #configureJaas(Resource) method.
cas.monitor.endpoints.ldap.base-dnBase DN to use.
Base DN to use. There may be scenarios where different parts of a single LDAP tree could be considered as base-dns. Rather than duplicating the LDAP configuration block for each individual base-dn, each entry can be specified and joined together using a special delimiter character. The user DN is retrieved using the combination of all base-dn and DN resolvers in the order defined. DN resolution should fail if multiple DNs are found. Otherwise the first DN found is returned. Usual syntax is: subtreeA,dc=example,dc=net|subtreeC,dc=example,dc=net.
cas.monitor.endpoints.ldap.bind-credentialThe bind credential to use when connecting to LDAP.
The bind credential to use when connecting to LDAP.
cas.monitor.endpoints.ldap.bind-dnThe bind DN to use when connecting to LDAP.
-
bindDn/bindCredentialprovided - Use the provided credentials to bind when initializing connections. -
bindDn/bindCredentialset to*- Use a fast-bind strategy to initialize the pool. -
bindDn/bindCredentialset to blank - Skip connection initializing; perform operations anonymously. - SASL mechanism provided - Use the given SASL mechanism to bind when initializing connections.
cas.monitor.endpoints.ldap.ldap-urlThe LDAP url to the server.
The LDAP url to the server. More than one may be specified, separated by space and/or comma.
cas.monitor.endpoints.ldap.search-filterUser filter to use for searching.
User filter to use for searching. Syntax is cn={user} or cn={0}.
You may also provide an external groovy script in the syntax of file:/path/to/GroovyScript.groovy to fully build the final filter template dynamically.
cas.monitor.endpoints.ldap.typeThe authentication type.
AUTHENTICATED-
AD- Users authenticate withsAMAccountName. -
AUTHENTICATED- Manager bind/search type of authentication. If {@code} principalAttributePassword} is empty then a user simple bind is done to validate credentials. Otherwise the given attribute is compared with the givenprincipalAttributePasswordusing theSHAencrypted value of it. -
ANONYMOUS: Similar semantics asAUTHENTICATEDexcept nobindDnandbindCredentialmay be specified to initialize the connection. IfprincipalAttributePasswordis empty then a user simple bind is done to validate credentials. Otherwise the given attribute is compared with the givenprincipalAttributePasswordusing theSHAencrypted value of it. - DIRECT: Direct Bind - Compute user DN from format string and perform simple bind. This is relevant when no search is required to compute the DN needed for a bind operation. Use cases for this type are: 1) All users are under a single branch in the directory,
e.g. ou=Users,dc=example,dc=org.2) The username provided on the CAS login form is part of the DN, e.g.uid=%s,ou=Users,dc=example,dc=org.
-
AD: Active Directory. -
AUTHENTICATED: Authenticated Search. -
DIRECT: Direct Bind. -
ANONYMOUS: Anonymous Search.
cas.monitor.endpoints.ldap.allow-multiple-dnsWhether search/query results are allowed to match on multiple DNs, or whether a single unique DN is expected for the result.
falseWhether search/query results are allowed to match on multiple DNs, or whether a single unique DN is expected for the result.
cas.monitor.endpoints.ldap.allow-multiple-entriesSet if multiple Entries are allowed.
falseSet if multiple Entries are allowed.
cas.monitor.endpoints.ldap.binary-attributesIndicate the collection of attributes that are to be tagged and processed as binary attributes by the underlying search resolver.
Indicate the collection of attributes that are to be tagged and processed as binary attributes by the underlying search resolver.
cas.monitor.endpoints.ldap.block-wait-timeThe length of time the pool will block.
PT3SThe length of time the pool will block. By default the pool will block indefinitely and there is no guarantee that waiting threads will be serviced in the order in which they made their request. This option should be used with a blocking connection pool when you need to control the exact number of connections that can be created
cas.monitor.endpoints.ldap.connect-timeoutSets the maximum amount of time that connects will block.
PT5SSets the maximum amount of time that connects will block.
cas.monitor.endpoints.ldap.connection-strategyIf multiple URLs are provided as the ldapURL this describes how each URL will be processed.
-
ACTIVE_PASSIVEFirst LDAP will be used for every request unless it fails and then the next shall be used. -
ROUND_ROBINFor each new connection the next url in the list will be used. -
RANDOMFor each new connection a random LDAP url will be selected. -
DNS_SRVLDAP urls based on DNS SRV records of the configured/given LDAP url will be used.
cas.monitor.endpoints.ldap.deref-aliasesDefine how aliases are de-referenced.
NEVER-
SEARCHING: dereference when searching the entries beneath the starting point but not when searching for the starting entry. -
FINDING: dereference when searching for the starting entry but not when searching the entries beneath the starting point. -
ALWAYS: dereference when searching for the starting entry and when searching the entries beneath the starting point.
cas.monitor.endpoints.ldap.disable-poolingWhether to use a pooled connection factory in components.
falseWhether to use a pooled connection factory in components.
cas.monitor.endpoints.ldap.dn-formatSpecify the dn format accepted by the AD authenticator, etc.
Specify the dn format accepted by the AD authenticator, etc. Example format might be uid=%s,ou=people,dc=example,dc=org.
cas.monitor.endpoints.ldap.enhance-with-entry-resolverWhether specific search entry resolvers need to be set on the authenticator, or the default should be used.
trueWhether specific search entry resolvers need to be set on the authenticator, or the default should be used.
cas.monitor.endpoints.ldap.fail-fastAttempt to populate the connection pool early on startup and fail quickly if something goes wrong.
trueAttempt to populate the connection pool early on startup and fail quickly if something goes wrong.
cas.monitor.endpoints.ldap.follow-referralsSet if search referrals should be followed.
trueSet if search referrals should be followed.
cas.monitor.endpoints.ldap.hostname-verifierHostname verification options.
DEFAULT-
DEFAULT: Default option, forcing verification. -
ANY: Skip hostname verification and allow all.
cas.monitor.endpoints.ldap.idle-timeRemoves connections from the pool based on how long they have been idle in the available queue.
PT10MRemoves connections from the pool based on how long they have been idle in the available queue. Prunes connections that have been idle for more than the indicated amount.
cas.monitor.endpoints.ldap.keystorePath to the keystore used for SSL connections.
Path to the keystore used for SSL connections. Typically contains SSL certificates for the LDAP server.
cas.monitor.endpoints.ldap.keystore-passwordKeystore password.
Keystore password.
cas.monitor.endpoints.ldap.keystore-typeThe type of keystore.
The type of keystore. PKCS12 or JKS. If left blank, defaults to the default keystore type indicated by the underlying Java platform.
cas.monitor.endpoints.ldap.ldap-authz.allow-multiple-resultsIndicate whether the LDAP search query is allowed to return multiple entries.
falseIndicate whether the LDAP search query is allowed to return multiple entries.
cas.monitor.endpoints.ldap.ldap-authz.base-dnBase DN to start the search.
Base DN to start the search.
cas.monitor.endpoints.ldap.ldap-authz.group-attributeAttribute expected to be found on the entry resulting from the group search whose value is going to be used to construct roles.
Attribute expected to be found on the entry resulting from the group search whose value is going to be used to construct roles. The final value is always prefixed with #groupPrefix. This is useful in scenarios where you wish to grant access to a resource to all users who a member of a given group.
cas.monitor.endpoints.ldap.ldap-authz.group-base-dnBase DN to start the search looking for groups.
Base DN to start the search looking for groups.
cas.monitor.endpoints.ldap.ldap-authz.group-filterSearch filter to begin looking for groups.
Search filter to begin looking for groups.
cas.monitor.endpoints.ldap.ldap-authz.group-prefixA prefix that is prepended to the group attribute value to construct an authorized role.
A prefix that is prepended to the group attribute value to construct an authorized role.
cas.monitor.endpoints.ldap.ldap-authz.role-attributeAttribute expected to be found on the entry whose value is going to be used to construct roles.
uugidAttribute expected to be found on the entry whose value is going to be used to construct roles. The final value is always prefixed with #rolePrefix. This is useful in scenarios where you wish to grant access to a resource to all users who carry a special attribute.
cas.monitor.endpoints.ldap.ldap-authz.role-prefixPrefix for the role.
ROLE_Prefix for the role.
cas.monitor.endpoints.ldap.ldap-authz.search-filterLDAP search filter to locate accounts.
LDAP search filter to locate accounts.
cas.monitor.endpoints.ldap.max-pool-sizeMaximum LDAP connection pool size which the pool can use to grow.
10Maximum LDAP connection pool size which the pool can use to grow.
cas.monitor.endpoints.ldap.min-pool-sizeMinimum LDAP connection pool size.
3Minimum LDAP connection pool size. Size the pool should be initialized to and pruned to
cas.monitor.endpoints.ldap.nameName of the LDAP handler.
Name of the LDAP handler.
cas.monitor.endpoints.ldap.page-sizeRequest that the server return results in batches of a specific size.
0Request that the server return results in batches of a specific size. See RFC 2696. This control is often used to work around server result size limits. A negative/zero value disables paged requests.
cas.monitor.endpoints.ldap.pool-passivatorYou may receive unexpected LDAP failures, when CAS is configured to authenticate using DIRECT or AUTHENTICATED types and LDAP is locked down to not allow anonymous binds/searches.
BINDDIRECT or AUTHENTICATED types and LDAP is locked down to not allow anonymous binds/searches. Every second attempt with a given LDAP connection from the pool would fail if it was on the same connection as a failed login attempt, and the regular connection validator would similarly fail. When a connection is returned back to a pool, it still may contain the principal and credentials from the previous attempt. Before the next bind attempt using that connection, the validator tries to validate the connection again but fails because it’s no longer trying with the configured bind credentials but with whatever user DN was used in the previous step. Given the validation failure, the connection is closed and CAS would deny access by default. Passivators attempt to reconnect to LDAP with the configured bind credentials, effectively resetting the connection to what it should be after each bind request. Furthermore if you are seeing errors in the logs that resemble a 'Operation exception encountered, reopening connection' type of message, this usually is an indication that the connection pool’s validation timeout established and created by CAS is greater than the timeout configured in the LDAP server, or more likely, in the load balancer in front of the LDAP servers. You can adjust the LDAP server session’s timeout for connections, or you can teach CAS to use a validity period that is equal or less than the LDAP server session’s timeout. Accepted values are: -
NONE: No passivation takes place. -
BIND: The default behavior which passivates a connection by performing a bind operation on it. This option requires the availability of bind credentials when establishing connections to LDAP.
cas.monitor.endpoints.ldap.principal-attribute-passwordIf principalAttributePassword is empty then a user simple bind is done to validate credentials otherwise the given attribute is compared with the given principalAttributePassword using the SHA encrypted value of it.
If principalAttributePassword is empty then a user simple bind is done to validate credentials otherwise the given attribute is compared with the given principalAttributePassword using the SHA encrypted value of it.
For the anonymous authentication type, if principalAttributePassword is empty then a user simple bind is done to validate credentials otherwise the given attribute is compared with the given principalAttributePassword using the SHA encrypted value of it.
cas.monitor.endpoints.ldap.prune-periodRemoves connections from the pool based on how long they have been idle in the available queue.
PT2HRemoves connections from the pool based on how long they have been idle in the available queue. Run the pruning process at the indicated interval.
cas.monitor.endpoints.ldap.resolve-from-attributeIf this attribute is set, the value found in the first attribute value will be used in place of the DN.
If this attribute is set, the value found in the first attribute value will be used in place of the DN.
cas.monitor.endpoints.ldap.response-timeoutDuration of time to wait for responses.
PT5SDuration of time to wait for responses.
cas.monitor.endpoints.ldap.sasl-authorization-idSASL authorization id.
SASL authorization id.
cas.monitor.endpoints.ldap.sasl-mechanismThe SASL mechanism.
The SASL mechanism.
cas.monitor.endpoints.ldap.sasl-mutual-authWhether SASL mutual authentication is enabled.
Whether SASL mutual authentication is enabled.
cas.monitor.endpoints.ldap.sasl-quality-of-protectionSASL quality of protected.
SASL quality of protected.
cas.monitor.endpoints.ldap.sasl-realmThe SASL realm.
The SASL realm.
cas.monitor.endpoints.ldap.sasl-security-strengthSASL security strength.
SASL security strength.
cas.monitor.endpoints.ldap.search-entry-handlersSearch handlers.
Search handlers.
cas.monitor.endpoints.ldap.search-entry-handlers[0].case-change.attribute-name-case-changeThe Attribute name case change.
The Attribute name case change.
cas.monitor.endpoints.ldap.search-entry-handlers[0].case-change.attribute-namesThe Attribute names.
The Attribute names.
cas.monitor.endpoints.ldap.search-entry-handlers[0].case-change.attribute-value-case-changeThe Attribute value case change.
The Attribute value case change.
cas.monitor.endpoints.ldap.search-entry-handlers[0].case-change.dn-case-changeThe Dn case change.
The Dn case change.
cas.monitor.endpoints.ldap.search-entry-handlers[0].dn-attribute.add-if-existsThe Add if exists.
The Add if exists.
cas.monitor.endpoints.ldap.search-entry-handlers[0].dn-attribute.dn-attribute-nameThe Dn attribute name.
entryDNThe Dn attribute name.
cas.monitor.endpoints.ldap.search-entry-handlers[0].merge-attribute.attribute-namesThe Attribute names.
The Attribute names.
cas.monitor.endpoints.ldap.search-entry-handlers[0].merge-attribute.merge-attribute-nameThe Merge attribute name.
The Merge attribute name.
cas.monitor.endpoints.ldap.search-entry-handlers[0].primary-group-id.base-dnThe Base dn.
The Base dn.
cas.monitor.endpoints.ldap.search-entry-handlers[0].primary-group-id.group-filterThe Group filter.
(&(objectClass=group)(objectSid={0}))The Group filter.
cas.monitor.endpoints.ldap.search-entry-handlers[0].recursive.merge-attributesThe Merge attributes.
The Merge attributes.
cas.monitor.endpoints.ldap.search-entry-handlers[0].recursive.search-attributeThe Search attribute.
The Search attribute.
cas.monitor.endpoints.ldap.search-entry-handlers[0].search-referral.limitThe default referral limit.
10The default referral limit.
cas.monitor.endpoints.ldap.search-entry-handlers[0].search-result.limitThe default referral limit.
10The default referral limit.
cas.monitor.endpoints.ldap.search-entry-handlers[0].typeThe type of search entry handler to choose.
-
FOLLOW_SEARCH_REFERRAL: Provides handling of an ldap referral for search operations. -
FOLLOW_SEARCH_RESULT_REFERENCE: Provides handling of an ldap continuation reference for search operations. -
ACTIVE_DIRECTORY: Process the entry results fetched from active directory and check for account status controls for disabled/expired accounts, etc. -
OBJECT_GUID: Object guid search entry handler. Handles theobjectGUIDattribute fetching and conversion. -
OBJECT_SID: Object sid search entry handler. Handles theobjectSidattribute fetching and conversion. -
CASE_CHANGE: Case change search entry handler. Provides the ability to modify the case of search entry DNs, attribute names, and attribute values. -
DN_ATTRIBUTE_ENTRY: DN attribute entry handler. Adds the entry DN as an attribute to the result set. Provides a client side implementation of RFC 5020. -
MERGE: Merge search entry handler. Merges the values of one or more attributes into a single attribute. -
PRIMARY_GROUP: Primary group search handler. Constructs the primary group SID and then searches for that group and puts it's DN in thememberOfattribute of the original search entry. -
RANGE_ENTRY: Range entry search handler. Rewrites attributes returned from Active Directory to include all values by performing additional searches. -
RECURSIVE_ENTRY: Recursive entry search handler. This recursively searches based on a supplied attribute and merges those results into the original entry. -
MERGE_ENTRIES: Merge entries handler. Merges the values of one or more attributes in all entries into a single attribute. The merged attribute may or may not already exist on the entry. If it does exist it's existing values will remain intact.
cas.monitor.endpoints.ldap.subtree-searchWhether subtree searching is allowed.
trueWhether subtree searching is allowed.
cas.monitor.endpoints.ldap.trust-certificatesPath of the trust certificates to use for the SSL connection.
Path of the trust certificates to use for the SSL connection. Ignores keystore-related settings when activated and used.
cas.monitor.endpoints.ldap.trust-managerTrust Manager options.
-
DEFAULT: Enable and force the default JVM trust managers. -
ANY: Trust any client or server.
cas.monitor.endpoints.ldap.trust-storePath to the keystore used to determine which certificates or certificate authorities should be trusted.
Path to the keystore used to determine which certificates or certificate authorities should be trusted. Used when connecting to an LDAP server via LDAPS or startTLS connection. If left blank, the default truststore for the Java runtime is used.
cas.monitor.endpoints.ldap.trust-store-passwordPassword needed to open the truststore.
Password needed to open the truststore.
cas.monitor.endpoints.ldap.trust-store-typeThe type of trust keystore that determines which certificates or certificate authorities are trusted.
The type of trust keystore that determines which certificates or certificate authorities are trusted. Types depend on underlying java platform, typically PKCS12 or JKS. If left blank, defaults to the default keystore type indicated by the underlying Java platform.
cas.monitor.endpoints.ldap.use-start-tlsWhether TLS should be used and enabled when establishing the connection.
falseWhether TLS should be used and enabled when establishing the connection.
cas.monitor.endpoints.ldap.validate-on-checkoutWhether connections should be validated when loaned out from the pool.
trueWhether connections should be validated when loaned out from the pool.
cas.monitor.endpoints.ldap.validate-periodPeriod at which pool should be validated.
PT5MPeriod at which pool should be validated.
cas.monitor.endpoints.ldap.validate-periodicallyWhether connections should be validated periodically when the pool is idle.
trueWhether connections should be validated periodically when the pool is idle.
cas.monitor.endpoints.ldap.validate-timeoutPeriod at which validation operations may time out.
PT5SPeriod at which validation operations may time out.
cas.monitor.endpoints.ldap.validator.attribute-nameAttribute name to use for the compare validator.
objectClassAttribute name to use for the compare validator.
cas.monitor.endpoints.ldap.validator.attribute-valueAttribute values to use for the compare validator.
topAttribute values to use for the compare validator.
cas.monitor.endpoints.ldap.validator.base-dnBase DN to use for the search request of the search validator.
Base DN to use for the search request of the search validator.
cas.monitor.endpoints.ldap.validator.dnDN to compare to use for the compare validator.
DN to compare to use for the compare validator.
cas.monitor.endpoints.ldap.validator.scopeSearch scope to use for the search request of the search validator.
OBJECTSearch scope to use for the search request of the search validator.
cas.monitor.endpoints.ldap.validator.search-filterSearch filter to use for the search request of the search validator.
(objectClass=*)Search filter to use for the search request of the search validator.
cas.monitor.endpoints.ldap.validator.typeDetermine the LDAP validator type.
searchDetermine the LDAP validator type.
-
search: Validates a connection is healthy by performing a search operation. Validation is considered successful if the search result size is greater than zero. -
none: No validation takes place. -
compare: Validates a connection is healthy by performing a compare operation.
LDAP & Active Directory
LDAP Scriptable Search Filter
LDAP search filters can point to an external Groovy script to dynamically construct the final filter template.
The script itself may be designed as:
1
2
3
4
5
6
7
8
9
import org.ldaptive.*
import org.springframework.context.*
def run(Object[] args) {
def (filter,parameters,applicationContext,logger) = args
logger.info("Configuring LDAP filter")
filter.setFilter("uid=something")
}
The following parameters are passed to the script:
| Parameter | Description |
|---|---|
filter |
FilterTemplate to be updated by the script and used for the LDAP query. |
parameters |
Map of query parameters which may be used to construct the final filter. |
applicationContext |
Reference to the Spring ApplicationContext reference. |
logger |
The object responsible for issuing log messages such as logger.info(...). |
To prepare CAS to support and integrate with Apache Groovy, please review this guide.
cas.monitor.endpoints.jdbc.driver-classThe JDBC driver used to connect to the database.
org.hsqldb.jdbcDriverThe JDBC driver used to connect to the database.
cas.monitor.endpoints.jdbc.passwordThe database connection password.
The database connection password.
cas.monitor.endpoints.jdbc.password-encoder.encoding-algorithmThe encoding algorithm to use such as MD5 .
The encoding algorithm to use such as MD5. Relevant when the type used is DEFAULT or GLIBC_CRYPT. When used with PasswordEncoderTypes#PBKDF2, it should be one of PBKDF2WithHmacSHA1, PBKDF2WithHmacSHA256 or PBKDF2WithHmacSHA512.
cas.monitor.endpoints.jdbc.password-encoder.typeDefine the password encoder type to use.
NONEDefine the password encoder type to use. Type may be specified as blank or NONE to disable password encoding. It may also refer to a fully-qualified class name that implements the Spring Security's PasswordEncoder interface if you wish you define your own encoder.
-
NONE: No password encoding (i.e. plain-text) takes place. -
DEFAULT: Use theDefaultPasswordEncoderof CAS. For message-digest algorithms viacharacter-encodingandencoding-algorithm. -
BCRYPT: Use theBCryptPasswordEncoderbased on the strength provided and an optional secret. -
SCRYPT: Use theSCryptPasswordEncoder. -
PBKDF2: Use thePbkdf2PasswordEncoderbased on the strength provided and an optional secret. -
STANDARD: Use theStandardPasswordEncoderbased on the secret provided. -
SSHA: Use theLdapShaPasswordEncodersupports Ldap SHA and SSHA (salted-SHA). The values are base-64 encoded and have the label {SHA} or {SSHA} prepended to the encoded hash. -
GLIBC_CRYPT: Use theGlibcCryptPasswordEncoderbased on theencoding-algorithm, strength provided and an optional secret. -
org.example.MyEncoder: An implementation ofPasswordEncoderof your own choosing. -
file:///path/to/script.groovy: Path to a Groovy script charged with handling password encoding operations.
cas.monitor.endpoints.jdbc.urlThe database connection URL.
jdbc:hsqldb:mem:cas-hsql-databaseThe database connection URL.
cas.monitor.endpoints.jdbc.userThe database user.
saThe database user.
The database user must have sufficient permissions to be able to handle schema changes and updates, when needed.
cas.monitor.endpoints.jdbc.autocommitThe default auto-commit behavior of connections in the pool.
falseThe default auto-commit behavior of connections in the pool. Determined whether queries such as update/insert should be immediately executed without waiting for an underlying transaction.
cas.monitor.endpoints.jdbc.batch-sizeA non-zero value enables use of JDBC2 batch updates by Hibernate. e.g. recommended values between 5 and 30.
100A non-zero value enables use of JDBC2 batch updates by Hibernate. e.g. recommended values between 5 and 30.
cas.monitor.endpoints.jdbc.connection-timeoutIndicates the maximum number of milliseconds that the service can wait to obtain a connection.
PT30SIndicates the maximum number of milliseconds that the service can wait to obtain a connection.
cas.monitor.endpoints.jdbc.data-source-nameAttempts to do a JNDI data source look up for the data source name specified.
Attempts to do a JNDI data source look up for the data source name specified. Will attempt to locate the data source object as is.
cas.monitor.endpoints.jdbc.ddl-autoHibernate feature to automatically validate and exports DDL to the schema.
updatevalidate or none may be more desirable for production, but any of the following options can be used: -
validate: Validate the schema, but make no changes to the database. -
update: Update the schema. -
create: Create the schema, destroying previous data. -
create-drop: Drop the schema at the end of the session. -
none: Do nothing.
Note that during a version migration where any schema has changed create-drop will result in the loss of all data as soon as CAS is started. For transient data like tickets this is probably not an issue, but in cases like the audit table important data could be lost. Using `update`, while safe for data, is confirmed to result in invalid database state. validate or none settings are likely the only safe options for production use.
For more info, see this.
cas.monitor.endpoints.jdbc.default-catalogQualifies unqualified table names with the given catalog in generated SQL.
Qualifies unqualified table names with the given catalog in generated SQL.
cas.monitor.endpoints.jdbc.default-schemaQualify unqualified table names with the given schema/tablespace in generated SQL.
Qualify unqualified table names with the given schema/tablespace in generated SQL.
cas.monitor.endpoints.jdbc.dialectThe database dialect is a configuration setting for platform independent software (JPA, Hibernate, etc) which allows such software to translate its generic SQL statements into vendor specific DDL, DML.
org.hibernate.dialect.HSQLDialectThe database dialect is a configuration setting for platform independent software (JPA, Hibernate, etc) which allows such software to translate its generic SQL statements into vendor specific DDL, DML.
cas.monitor.endpoints.jdbc.fail-fast-timeoutSet the pool initialization failure timeout.
1- Any value greater than zero will be treated as a timeout for pool initialization. The calling thread will be blocked from continuing until a successful connection to the database, or until the timeout is reached. If the timeout is reached, then a
PoolInitializationExceptionwill be thrown. - A value of zero will not prevent the pool from starting in the case that a connection cannot be obtained. However, upon start the pool will attempt to obtain a connection and validate that the
connectionTestQueryandconnectionInitSqlare valid. If those validations fail, an exception will be thrown. If a connection cannot be obtained, the validation is skipped and the pool will start and continue to try to obtain connections in the background. This can mean that callers toDataSource#getConnection()may encounter exceptions. - A value less than zero will not bypass any connection attempt and validation during startup, and therefore the pool will start immediately. The pool will continue to try to obtain connections in the background. This can mean that callers to
DataSource#getConnection()may encounter exceptions.
connectionTimeout or validationTimeout; they will be honored before this timeout is applied. The default value is one millisecond.cas.monitor.endpoints.jdbc.fetch-sizeUsed to specify number of rows to be fetched in a select query.
100Used to specify number of rows to be fetched in a select query.
cas.monitor.endpoints.jdbc.generate-statisticsAllow hibernate to generate query statistics.
falseAllow hibernate to generate query statistics.
cas.monitor.endpoints.jdbc.health-queryThe SQL query to be executed to test the validity of connections.
The SQL query to be executed to test the validity of connections. This is for "legacy" databases that do not support the JDBC4 Connection.isValid() API.
cas.monitor.endpoints.jdbc.idle-timeoutControls the maximum amount of time that a connection is allowed to sit idle in the pool.
PT10MControls the maximum amount of time that a connection is allowed to sit idle in the pool.
cas.monitor.endpoints.jdbc.isolate-internal-queriesThis property determines whether data source isolates internal pool queries, such as the connection alive test, in their own transaction.
falseThis property determines whether data source isolates internal pool queries, such as the connection alive test, in their own transaction.
Since these are typically read-only queries, it is rarely necessary to encapsulate them in their own transaction. This property only applies if #autocommit is disabled.
cas.monitor.endpoints.jdbc.isolation-level-nameDefines the isolation level for transactions.
ISOLATION_READ_COMMITTEDDefines the isolation level for transactions. @see org.springframework.transaction.TransactionDefinition
cas.monitor.endpoints.jdbc.leak-thresholdControls the amount of time that a connection can be out of the pool before a message is logged indicating a possible connection leak.
PT6SControls the amount of time that a connection can be out of the pool before a message is logged indicating a possible connection leak.
cas.monitor.endpoints.jdbc.password-encoder.character-encodingThe encoding algorithm to use such as 'UTF-8'.
UTF-8The encoding algorithm to use such as 'UTF-8'. Relevant when the type used is DEFAULT.
cas.monitor.endpoints.jdbc.password-encoder.hash-lengthWhen used by PasswordEncoderTypes#ARGON2 , it indicates the hash strength/length.
16When used by PasswordEncoderTypes#ARGON2, it indicates the hash strength/length.
cas.monitor.endpoints.jdbc.password-encoder.iterationsWhen used by PasswordEncoderTypes#PBKDF2 , it indicates the required number of iterations.
310000When used by PasswordEncoderTypes#PBKDF2, it indicates the required number of iterations.
cas.monitor.endpoints.jdbc.password-encoder.secretSecret to use with PasswordEncoderTypes#STANDARD , PasswordEncoderTypes#PBKDF2 , PasswordEncoderTypes#BCRYPT , PasswordEncoderTypes#GLIBC_CRYPT password encoders.
Secret to use with PasswordEncoderTypes#STANDARD, PasswordEncoderTypes#PBKDF2, PasswordEncoderTypes#BCRYPT, PasswordEncoderTypes#GLIBC_CRYPT password encoders. Secret usually is an optional setting.
cas.monitor.endpoints.jdbc.password-encoder.strengthStrength or number of iterations to use for password hashing.
16Strength or number of iterations to use for password hashing. Usually relevant when dealing with PasswordEncoderTypes#BCRYPT, PasswordEncoderTypes#PBKDF2 or PasswordEncoderTypes#GLIBC_CRYPT. When used by PasswordEncoderTypes#ARGON2 or PasswordEncoderTypes#PBKDF2, it indicates the salt strength.
cas.monitor.endpoints.jdbc.physical-naming-strategy-class-nameFully-qualified name of the class that can control the physical naming strategy of hibernate.
org.apereo.cas.hibernate.CasHibernatePhysicalNamingStrategyFully-qualified name of the class that can control the physical naming strategy of hibernate.
cas.monitor.endpoints.jdbc.pool.keep-alive-timeThis property controls the keepalive interval for a connection in the pool.
0This property controls the keepalive interval for a connection in the pool. An in-use connection will never be tested by the keepalive thread, only when it is idle will it be tested. Default is zero, which disables this feature.
cas.monitor.endpoints.jdbc.pool.max-sizeControls the maximum number of connections to keep in the pool, including both idle and in-use connections.
18Controls the maximum number of connections to keep in the pool, including both idle and in-use connections.
cas.monitor.endpoints.jdbc.pool.max-waitSets the maximum time in seconds that this data source will wait while attempting to connect to a database.
PT2SSets the maximum time in seconds that this data source will wait while attempting to connect to a database.
A value of zero specifies that the timeout is the default system timeout if there is one; otherwise, it specifies that there is no timeout.
cas.monitor.endpoints.jdbc.pool.maximum-lifetimeThis property controls the maximum lifetime of a connection in the pool.
PT10MThis property controls the maximum lifetime of a connection in the pool. When a connection reaches this timeout, even if recently used, it will be retired from the pool. An in-use connection will never be retired, only when it is idle will it be removed.
cas.monitor.endpoints.jdbc.pool.min-sizeControls the minimum size that the pool is allowed to reach, including both idle and in-use connections.
6Controls the minimum size that the pool is allowed to reach, including both idle and in-use connections.
cas.monitor.endpoints.jdbc.pool.nameSet the name of the connection pool.
Set the name of the connection pool. This is primarily used for the MBean to uniquely identify the pool configuration.
cas.monitor.endpoints.jdbc.pool.suspensionWhether or not pool suspension is allowed.
falseWhether or not pool suspension is allowed.
There is a performance impact when pool suspension is enabled. Unless you need it (for a redundancy system for example) do not enable it.
cas.monitor.endpoints.jdbc.pool.timeout-millisThe maximum number of milliseconds that the pool will wait for a connection to be validated as alive.
1000The maximum number of milliseconds that the pool will wait for a connection to be validated as alive.
cas.monitor.endpoints.jdbc.propagation-behavior-nameDefines the propagation behavior for transactions.
PROPAGATION_REQUIREDDefines the propagation behavior for transactions. @see org.springframework.transaction.TransactionDefinition
cas.monitor.endpoints.jdbc.propertiesAdditional settings provided by Hibernate (or the connection provider) in form of key-value pairs.
Additional settings provided by Hibernate (or the connection provider) in form of key-value pairs.
cas.monitor.endpoints.jdbc.queryQuery to execute in order to authenticate users via JDBC.
Query to execute in order to authenticate users via JDBC. Example: SELECT username,password,enabled FROM users WHERE username=?
cas.monitor.endpoints.jdbc.read-onlyConfigures the Connections to be added to the pool as read-only Connections.
falseConfigures the Connections to be added to the pool as read-only Connections.
cas.monitor.endpoints.jdbc.role-prefixPrefix to add to the role.
Prefix to add to the role.
Hibernate & JDBC
Control global properties that are relevant to Hibernate, when CAS attempts to employ and utilize database resources, connections and queries.
cas.jdbc.gen-ddlWhether to generate DDL after the EntityManagerFactory has been initialized creating/updating all relevant tables.
trueWhether to generate DDL after the EntityManagerFactory has been initialized creating/updating all relevant tables.
cas.jdbc.physical-table-namesIndicate a physical table name to be used by the hibernate naming strategy in case table names need to be customized for the specific type of database.
Indicate a physical table name to be used by the hibernate naming strategy in case table names need to be customized for the specific type of database. The key here indicates the CAS-provided table name and the value is the translate physical name for the database. If a match is not found for the CAS-provided table name, then that name will be used by default.
cas.jdbc.show-sqlWhether SQL queries should be displayed in the console/logs.
falseWhether SQL queries should be displayed in the console/logs.
Password encoding
If you need to design your own password encoding scheme where the type is specified as a fully qualified Java class name, the structure of the class would be similar to the following:
1
2
3
4
5
6
7
8
9
10
11
package org.example.cas;
import org.springframework.security.crypto.codec.*;
import org.springframework.security.crypto.password.*;
public class MyEncoder extends AbstractPasswordEncoder {
@Override
protected byte[] encode(CharSequence rawPassword, byte[] salt) {
return ...
}
}
If you need to design your own password encoding scheme where the type is specified as a path to a Groovy script, the structure of the script would be similar to the following:
1
2
3
4
5
6
7
8
9
10
11
12
import java.util.*
byte[] run(final Object... args) {
def (rawPassword,generatedSalt,logger,applicationContext) = args
logger.debug("Encoding password...")
return ...
}
Boolean matches(final Object... args) {
def (rawPassword,encodedPassword,logger,applicationContext) = args
logger.debug("Does match or not ?");
return ...
To prepare CAS to support and integrate with Apache Groovy, please review this guide.
A static list of users, passwords and roles in a JSON file.
Supported password encodings are {sha512}, {sha256}, {bcrypt},
{noop}, {pbkdf2}, {scrypt} and {argon2}.
1
2
3
4
5
6
7
[
{
"username": "casuser",
"password": "{sha512}<hashed-password>",
"authorities": [ "ROLE_ADMIN" ]
}
]
cas.monitor.endpoints.json.locationThe location of the resource.
The location of the resource. Resources can be URLs, or files found either on the classpath or outside somewhere in the file system.
In the event the configured resource is a Groovy script, especially if the script is set to reload on changes, you may need to adjust the total number of inotify instances. On Linux, you may need to add the following line to /etc/sysctl.conf: fs.inotify.max_user_instances = 256.
You can check the current value via cat /proc/sys/fs/inotify/max_user_instances.
In situations and scenarios where CAS is able to automatically watch the underlying resource for changes and detect updates and modifications dynamically, you may be able to specify the following setting as either an environment variable or system property with a value of false to disable the resource watcher: org.apereo.cas.util.io.PathWatcherService.
cas.monitor.endpoints.endpoint.oidcFederationJwks.accessDefine the security access level of the endpoint.
DENY-
PERMIT: Allow open access to the endpoint. -
ANONYMOUS: Allow anonymous access to the endpoint. -
DENY: Block access to the endpoint. -
AUTHENTICATED: Require authenticated access to the endpoint. -
ROLE: Require authenticated access to the endpoint along with a role requirement. -
AUTHORITY: Require authenticated access to the endpoint along with an authority requirement. -
IP_ADDRESS: Require authenticated access to the endpoint using a collection of IP addresses.
cas.monitor.endpoints.endpoint.oidcFederationJwks.required-authoritiesRequired user authorities.
Required user authorities.
cas.monitor.endpoints.endpoint.oidcFederationJwks.required-ip-addressesRequired IP addresses.
Required IP addresses. CIDR ranges are accepted.
cas.monitor.endpoints.endpoint.oidcFederationJwks.required-rolesRequired user roles.
Required user roles.
cas.monitor.endpoints.form-login-enabledControl whether access to endpoints can be controlled via form-based login over the web via a special admin login endpoint.
falseControl whether access to endpoints can be controlled via form-based login over the web via a special admin login endpoint.
management.endpoint.health.accessPermitted level of access for the health endpoint.
unrestrictedPermitted level of access for the health endpoint.
management.endpoint.health.cache.time-to-liveMaximum time that a response can be cached.
0msMaximum time that a response can be cached.
management.endpoint.health.groupHealth endpoint groups.
Health endpoint groups.
management.endpoint.health.logging.slow-indicator-thresholdThreshold after which a warning will be logged for slow health indicators.
10sThreshold after which a warning will be logged for slow health indicators.
management.endpoint.health.probes.add-additional-pathsWhether to make the liveness and readiness health groups available on the main server port.
falseWhether to make the liveness and readiness health groups available on the main server port.
management.endpoint.health.probes.enabledWhether to enable liveness and readiness probes.
trueWhether to enable liveness and readiness probes.
management.endpoint.health.rolesRoles used to determine whether a user is authorized to be shown details.
Roles used to determine whether a user is authorized to be shown details. When empty, all authenticated users are authorized.
management.endpoint.health.show-componentsWhen to show components.
When to show components. If not specified the 'show-details' setting will be used.
management.endpoint.health.show-detailsWhen to show full health details.
neverWhen to show full health details.
management.endpoint.health.status.http-mappingMapping of health statuses to HTTP status codes.
Mapping of health statuses to HTTP status codes. By default, registered health statuses map to sensible defaults (for example, UP maps to 200).
management.endpoint.health.status.orderList of health statuses in order of severity.
["DOWN", "OUT_OF_SERVICE", "UP", "UNKNOWN"]List of health statuses in order of severity.
management.endpoint.health.validate-group-membershipWhether to validate health group membership on startup.
trueWhether to validate health group membership on startup. Validation fails if a group includes or excludes a health contributor that does not exist.
management.endpoints.access.defaultDefault access level for all endpoints.
Default access level for all endpoints.
management.endpoints.access.max-permittedMaximum level of endpoint access that is permitted.
unrestrictedMaximum level of endpoint access that is permitted. Caps an endpoint's individual access level (management.endpoint.<id>.access) and the default access (management.endpoints.access.default).'
management.endpoints.enabled-by-defaultWhether to enable or disable all endpoints by default.
Whether to enable or disable all endpoints by default.
management.endpoints.jackson.isolated-json-mapperWhether to use an isolated JsonMapper to serialize endpoint JSON.
trueWhether to use an isolated JsonMapper to serialize endpoint JSON.
management.endpoints.jackson2.isolated-object-mapperWhether to use an isolated object mapper to serialize endpoint JSON.
trueWhether to use an isolated object mapper to serialize endpoint JSON.
management.endpoints.jmx.domainEndpoints JMX domain name.
org.springframework.bootEndpoints JMX domain name. Fallback to 'spring.jmx.default-domain' if set.
management.endpoints.jmx.exposure.excludeEndpoint IDs that should be excluded or '*' for all.
Endpoint IDs that should be excluded or '*' for all.
management.endpoints.jmx.exposure.includeEndpoint IDs that should be included or '*' for all.
healthEndpoint IDs that should be included or '*' for all.
management.endpoints.jmx.static-namesAdditional static properties to append to all ObjectNames of MBeans representing Endpoints.
Additional static properties to append to all ObjectNames of MBeans representing Endpoints.
management.endpoints.jmx.unique-namesWhether unique runtime object names should be ensured.
Whether unique runtime object names should be ensured.
management.endpoints.migrate-legacy-idsWhether to transparently migrate legacy endpoint IDs.
falseWhether to transparently migrate legacy endpoint IDs.
management.endpoints.web.base-pathBase path for Web endpoints.
/actuatorBase path for Web endpoints. Relative to the servlet context path (server.servlet.context-path) or WebFlux base path (spring.webflux.base-path) when the management server is sharing the main server port. Relative to the management server base path (management.server.base-path) when a separate management server port (management.server.port) is configured.
management.endpoints.web.cors.allow-credentialsWhether credentials are supported.
Whether credentials are supported. When not set, credentials are not supported.
management.endpoints.web.cors.allowed-headersList of headers to allow in a request. '*' allows all headers.
List of headers to allow in a request. '*' allows all headers.
management.endpoints.web.cors.allowed-methodsList of methods to allow. '*' allows all methods.
List of methods to allow. '*' allows all methods. When not set, defaults to GET.
management.endpoints.web.cors.allowed-origin-patternsList of origin patterns to allow.
List of origin patterns to allow. Unlike allowed origins which only supports '*', origin patterns are more flexible (for example 'https://*.example.com') and can be used when credentials are allowed. When no allowed origin patterns or allowed origins are set, CORS support is disabled.
management.endpoints.web.cors.allowed-originsList of origins to allow. '*' allows all origins.
List of origins to allow. '*' allows all origins. When credentials are allowed, '*' cannot be used and origin patterns should be configured instead. When no allowed origins or allowed origin patterns are set, CORS support is disabled.
management.endpoints.web.cors.exposed-headersList of headers to include in a response.
List of headers to include in a response.
management.endpoints.web.cors.max-ageHow long the response from a pre-flight request can be cached by clients.
1800sHow long the response from a pre-flight request can be cached by clients. If a duration suffix is not specified, seconds will be used.
management.endpoints.web.discovery.enabledWhether the discovery page is enabled.
trueWhether the discovery page is enabled.
management.endpoints.web.exposure.excludeEndpoint IDs that should be excluded or '*' for all.
Endpoint IDs that should be excluded or '*' for all.
management.endpoints.web.exposure.includeEndpoint IDs that should be included or '*' for all.
["health"]Endpoint IDs that should be included or '*' for all.
management.endpoints.web.path-mappingMapping between endpoint IDs and the path that should expose them.
Mapping between endpoint IDs and the path that should expose them.
management.health.binders.enabledAllows to enable/disable binder's' health indicators.
trueAllows to enable/disable binder's' health indicators. If you want to disable health indicator completely, then set it to `false`.
management.health.db.enabledWhether to enable database health check.
trueWhether to enable database health check.
management.health.db.ignore-routing-data-sourcesWhether to ignore AbstractRoutingDataSources when creating database health indicators.
falseWhether to ignore AbstractRoutingDataSources when creating database health indicators.
management.health.defaults.enabledWhether to enable default health indicators.
trueWhether to enable default health indicators.
management.health.diskspace.enabledWhether to enable disk space health check.
trueWhether to enable disk space health check.
management.health.diskspace.pathPath used to compute the available disk space.
Path used to compute the available disk space.
management.health.diskspace.thresholdMinimum disk space that should be available.
10MBMinimum disk space that should be available.
management.health.influxdb.enabled
management.health.livenessstate.enabledWhether to enable liveness state health check.
falseWhether to enable liveness state health check.
management.health.mail.enabledWhether to enable Mail health check.
trueWhether to enable Mail health check.
management.health.mongo.enabled
management.health.mongodb.enabledWhether to enable MongoDB health check.
trueWhether to enable MongoDB health check.
management.health.ping.enabledWhether to enable ping health check.
trueWhether to enable ping health check.
management.health.probes.enabledWhether to enable liveness and readiness probes.
falseWhether to enable liveness and readiness probes.
management.health.pubsub.enabledWhether to enable the Pub/Sub health indicator when used with Spring Boot Actuator.
trueWhether to enable the Pub/Sub health indicator when used with Spring Boot Actuator.
management.health.rabbit.enabledWhether to enable RabbitMQ health check.
trueWhether to enable RabbitMQ health check.
management.health.readinessstate.enabledWhether to enable readiness state health check.
falseWhether to enable readiness state health check.
management.health.redis.enabledWhether to enable Redis health check.
trueWhether to enable Redis health check.
management.health.refresh.enabledEnable the health endpoint for the refresh scope.
trueEnable the health endpoint for the refresh scope.
management.health.ssl.certificate-validity-warning-thresholdIf an SSL Certificate will be invalid within the time span defined by this threshold, it should trigger a warning.
14dIf an SSL Certificate will be invalid within the time span defined by this threshold, it should trigger a warning.
management.health.ssl.enabledWhether to enable SSL certificate health check.
trueWhether to enable SSL certificate health check.
management.health.zookeeper.enabledEnable the health endpoint for zookeeper.
trueEnable the health endpoint for zookeeper.
-
404: the endpoint is not enabled or not exposed; check the Enable & expose tab. -
401or403: the access rule or the credentials rejected the call; check the Security tab.
For more detail, raise these log levels in the log4j configuration:
1
2
3
4
5
6
7
8
<Logger name="org.apereo.cas.oidc.federation.web" level="debug" additivity="false">
<AppenderRef ref="console" />
<AppenderRef ref="file" />
</Logger>
<Logger name="org.springframework.security" level="debug" additivity="false">
<AppenderRef ref="console" />
<AppenderRef ref="file" />
</Logger>
The keystore must be writable for rotation and revocation; CAS writes a new keystore file next to it, with the permissions, owner and group of the keystore where the file system allows it, and not readable by a group other than the keystore’s; a new keystore is readable by its owner only where the file system has such permissions. The new file then replaces the keystore at once, so that the keystore is never read half written. A keystore that is a symbolic link is written where the link points. Where the file system does not allow replacing the keystore while it is in use, as Windows may, the keys are not saved and stay as they were. CAS reads the keystore again when it changes, so nodes that share the keystore file sign with the same keys; nodes that do not share it each have their own keys, which superiors would all have to know. Superiors verify the entity configuration of CAS with the keys of their subordinate statement about it: when CAS is a subordinate, its future key should be added to the subordinate definition at each superior before the rotation makes it current. Likewise, entities that trust a CAS trust anchor with keys obtained out of band should be given its future key ahead of time.
Trust Anchors
An OpenID Provider accepts the trust chains of relying parties that end at one of its trust anchors. Each trust anchor is defined by its entity identifier and its keys, which should be obtained out of band, as OpenID Federation recommends, and may be given as a file, a URL or the JSON web key set itself; private key members are ignored. A trust anchor defined without keys is trusted with the keys of its own entity configuration, that is with whoever answers at its entity identifier. Trust anchors and their keys can be defined using CAS properties.
The authority hints are the entities directly above CAS, which issue subordinate statements about it; they are not trust anchors unless they are also defined as such. When no trust anchor is defined, CAS uses the authority hints as trust anchors, with the keys of their own entity configuration, and logs a warning.
Automatic Registration
As an OpenID Provider, CAS registers relying parties of the federation automatically, as described by OpenID Federation, without any prior registration step:
- The relying party uses its entity identifier as its client id, which must be an
httpsURL. On the first authorization or pushed authorization request, CAS resolves the trust chains of the relying party up to one of the trust anchors, and registers the resolvedopenid_relying_partymetadata as an application that expires with the trust chain and is resolved again shortly before. - Each redirect URI of the relying party is matched exactly, whatever characters it contains.
- The relying party authenticates with
private_key_jwt, assumed when its metadata names no method, or withtls_client_auth. A relying party whose metadata names any other method is not registered, and so is a relying party that publishes its keys at ajwks_urithat is not anhttpsURL. - Every authorization request proves that the relying party controls its keys. It either carries a request object
signed with a key of the relying party’s metadata, whose
issandclient_idare the entity identifier, whoseaudis only the issuer of CAS, withjtiandexpand withoutsub, valid for 10 minutes at most, or it refers withrequest_urito a pushed authorization request for which the relying party authenticated itself. Client assertions may be addressed to the issuer of CAS as well as to its token endpoint, except at the pushed authorization endpoint, where the client assertion of an automatically registered relying party must be addressed to the issuer of CAS only, as OpenID Federation requires; otherwise the request is refused with401andinvalid_client. - A request object is used once: CAS remembers its
jtifor as long as the request object could be accepted, and at least for the lifetime of transient session tickets, during which only the very same request object is accepted again, as happens when the browser returns from the login. - A request object may also be encrypted to CAS; it is decrypted first, and the signed request object inside is checked as above.
- Requests that fail these checks are refused with
invalid_requestand are never redirected to the relying party. - The relying party that must prove its keys is the client of the authorization request: the
client_idparameter or, when the request carries none, theclient_idof its request object, on the OpenID Connect and the OAuth authorization endpoints alike, including the OpenID Connect authorization and pushed authorization endpoints that also answer outside the OpenID Connect base path, whatever the encoding of the path. A request object whoseclient_iddiffers from theclient_idparameter is refused for every client, as OpenID Connect requires. - The constraints that trust anchors and intermediates set in their subordinate statements apply to every trust chain
of a relying party, whether resolved or sent, as
OpenID Federation requires: the maximum number of
intermediates below the entity that sets them, the permitted and excluded domain names of the entities below it, and
the entity types they may have, which must include
openid_relying_party. A trust chain that breaks any of them is not used. The trust chains of trust mark issuers are held to the same constraints, except for entity types. - The metadata of a relying party is resolved from its trust chain as
OpenID Federation describes: the metadata the
immediate superior states about the relying party takes precedence over the metadata the relying party publishes,
and the metadata policies of the trust anchor and intermediates are then merged, from the trust anchor down, and
applied.
scopeis handled as a list of scope values. Operators beyond the standard ones are ignored, unless a statement of the chain lists them as critical, in which case the trust chain is not used. A trust chain whose policies conflict, or whose policies the metadata does not satisfy, such as a missing essential parameter, is not used either. - The keys a trust anchor is trusted with out of band may still list a key the trust anchor has since revoked. When the
trust anchor publishes its historical keys, CAS reads them, verified with those keys, and does not use a trust chain
in which the trust anchor signed with a key it marks revoked, or with a key it had already retired when it issued the
statement. The historical keys are always found through the entity configuration CAS fetches from the trust anchor,
never through one a trust chain sent by a relying party carries; they must be served over
https, and each must carry its expiration, as the specification requires. A trust anchor that advertises no historical keys has revoked none. CAS keeps the historical keys of a trust anchor for a minute. When its entity configuration or the historical keys it advertises cannot be obtained or verified, CAS cannot tell whether the key was revoked and does not use the trust chain for now; the relying party is not remembered as unresolved on that account, so it can be registered as soon as the trust anchor can be reached again. Every statement of a trust chain must name the key it is signed with, as the specification requires; a trust chain with a statement that does not is not used. - The metadata superiors assert about a relying party, and their metadata policies, apply only to the entity types
its entity configuration publishes: an entity that publishes no
openid_relying_partymetadata is not registered, whatever its superiors state. - Resolving trust chains is bounded: CAS follows a limited number of authority hints from any entity configuration and
accepts a limited number of intermediates between the relying party and a trust anchor, and an entity identifier whose
trust chains could not be resolved is not resolved again for a while. Meanwhile, a registration that could not be
refreshed is kept as it is. Each resolution sends a limited number of requests, 50 by default, counting entity
configurations, subordinate statements, historical keys and trust mark status requests together, including those for
the issuers of its trust marks; a statement fetched once is reused within the same resolution, and a resolution that
needs more fails. The historical keys of trust anchors are not counted. Each resolution must also be done within a
limited time, 10 seconds by default, all of its requests together, however slowly the other side answers, and every
response is limited in size; the time limit covers each whole request, from the host name lookup and the TLS handshake
to the last byte of the response. Requests are only sent over
https, never to link-local, unspecified or multicast addresses such as the metadata service of a cloud provider, including such addresses embedded in IPv6 addresses, use the TLS settings of CAS and do not follow redirects. Addresses written in a URL are always checked, and host names when CAS resolves them, as it does to connect directly or through a SOCKS proxy; host names sent through an HTTP proxy of the JVM, or through a SOCKS proxy when CAS cannot resolve them, are resolved by the proxy, which then decides where they may go. A limited number of requests run at the same time, those abandoned at the deadline included. A trust anchor that does not accept a connection or answer in the time it is given, or that answers with a server error, is treated as unreachable. A resolution that runs out of time or requests, or finds no room among the running requests, cannot tell whether a relying party has a valid trust chain or trust mark: it is unavailable, and the relying party is not remembered as unresolved. A limited number of resolutions run at the same time, 10 by default, for automatic registration and the resolve endpoint together; one that cannot start within a second fails without the entity being remembered as unresolved, and so does a registration that cannot get its turn within a few seconds. Using the trust chain a relying party sent and, when it cannot be used, resolving its trust chains is a single resolution, within one time limit and request budget. A relying party that is not registered yet and cannot be registered for now is answered withtemporarily_unavailablerather than as an unknown client. These limits can be tuned using CAS properties. - A relying party whose metadata asks for pairwise subject identifiers receives them, calculated with a salt that can be
defined using CAS properties, so that they stay the same whenever the relying party is registered again and are
shared by relying parties of the same sector; the salt is never written into its registration. Without that salt, such
a relying party is not registered. Its sector must be its own: its sector identifier URI, when it has one, must be an
httpsURL, is fetched as a request of the resolution and must list all of its redirect URIs, and without one, its redirect URIs must all be in the same sector, read as they are when identifiers are calculated: anhttporhttpsredirect URI on a host other than a loopback one is in the sector of its host, and any other one in the relying party’s own. A sector identifier document that cannot be fetched for now leaves the relying party unregistered for now, without remembering it as unresolved, and its other trust anchors are tried. - A relying party may send its trust chain in the
trust_chainheader of its request object or client assertion, as OpenID Federation allows. Whenever CAS would resolve the relying party, to register it, to refresh its registration or after a key rollover, it uses that chain instead, once fully verified: the chain must start with the entity configuration of the relying party, end at a trust anchor whose keys are defined out of band, verify statement by statement from those keys, not have expired and not exceed the maximum path length; the required trust marks and the metadata policies of the chain apply as usual. Statements are public, so a chain alone does not show who sends it: it is used only when its entity configuration was not issued before the one the current registration was built from, and when the request object or client assertion that carries it is signed with the keys it resolves to. A chain replayed by someone else, even a genuine one, cannot keep or bring back keys the relying party has replaced. CAS resolves the trust chains itself when the sent chain cannot be used, and always when the trust anchor is trusted with the keys of its own entity configuration, which a sent chain could forge. - A relying party that rolls its keys over is resolved again before its registration expires, as OpenID Federation allows: when a request object or a client assertion it sends, at any endpoint, verifies with none of the keys of its registration, CAS resolves its trust chains first and then checks the request with the keys the relying party now publishes. Each relying party is resolved this way at most once in the same period that applies to entity identifiers that could not be resolved, so objects signed with keys a relying party never published cannot make CAS fetch its statements over and over. Such objects count against that period whoever sends them, and the relying party is then resolved from what it publishes, so a relying party should publish a new key before it signs with it.
Required Trust Marks
An OpenID Provider may only register relying parties that hold specific trust marks, such as a certification or a membership the federation vouches for. The required trust mark types can be defined using CAS properties. For each of them, the entity configuration of the relying party must carry a trust mark of that type that:
- is a signed JWT typed
trust-mark+jwtthat names its signing key, is about the relying party, and has been issued and has not expired; - is issued by an entity that the trust anchor of the relying party’s trust chain accepts for that type in its
trust_mark_issuers; - verifies with the federation keys of its issuer: the trust anchor’s own keys, or the keys of the issuer’s entity configuration, whose trust chain must end at the same trust anchor. A trust mark signed with a key the issuer has since retired verifies with that key as the issuer’s historical keys publish it, provided the key was never revoked and the trust mark was issued before the key was retired. The historical keys of an issuer are kept for a minute, except when they could not be requested: a resolution that has used up its requests verifies the trust mark with the current keys of its issuer only, and one that runs out of time or finds no room among the running requests is unavailable;
- carries a valid
delegationfrom the owner of the type, when the trust anchor names one intrust_mark_owners; - is reported
activeby the issuer’s trust mark status endpoint, when the issuer publishes one; a status endpoint or historical keys endpoint that cannot be reached for now or answers with a server error makes the resolution unavailable rather than the trust mark invalid, unless another trust mark of the type is valid, or the relying party lacks a valid trust mark of another required type; the endpoint must be anhttpsURL. So does a trust anchor that cannot be reached while the trust chain of the issuer is resolved. An issuer whose endpoints keep answering with server errors thus keeps the relying parties holding its trust marks from being remembered as unresolved, so they are resolved again with each request, each time within the limits of a resolution.
A relying party that lacks a valid trust mark of a required type is not registered, and the reason is reported in the logs. Trust marks are checked again whenever the registration is refreshed.
An authorization request passed by value travels through the CAS login as part of its service URL, and CAS does not
match service URLs beyond a maximum length, which can be adjusted using CAS properties. A request object that carries
its trust_chain is several kilobytes long: relying parties should send it with a
pushed authorization request, as OpenID Federation suggests, or the limit must be
raised.