required-ip-addresses: ["127.0.0.1", "172.18.0.0/22", "10.0.0.0/8"] don't work anymore. this undocumented breaking change caused deployment to fail because health endpoint cannot be accessed anymore
Dear Apereo CAS community, I am in need of some help if someone has come across a similar issue.
I'n on CAS 6.6.0 and I'm trying to implement SAML 2.0 integration by having CAS acting as a SP and delegating authentication to external IDP.
I have configured all settings to generate the SP metadata file, and I was provided with the IDP metadata file as well.
The problem is that the IDP entity wants to use the Artifact resolution service and sends us back a SAMLart (artifact). At this point, the flow returns back to CAS and it stays there.
It seems that CAS does not know what to do with the SAML artifact once it receives it from IDP, when technically it should be calling the endpoint specified by the ArtifactResolutionService from the IDP metadata file.
From tracing the SAML request/responses, I can tell that CAS is not calling anything else once it receives the SAMLart from the IDP after user authenticates successfully.
I ran a separate test by using the IDP service at https://samltest.id/ and everything is working fine, but I think the difference here is that this IDP doesn't use artifact binding, it goes a simpler way of HTTP-POST I believe.
So my question is: Does CAS SAML 2.0 integration support this artifact binding (sending artifact resolve request / receiving artifact response)?
If so, what could I be missing in my configurations?
If not, where would I need to modify CAS code to perform a SOAP request with the SAML artifact to receive the assertion?
1 reply
somehow
classpath:config/appliction.yamlis being ignored butclasspath:application.yamlgets loaded. according to spring boot 2.7 docsconfig/application.yamlshould be loaded as before. so what did change in cas 6.6 that changed this behavior?
i tracked down this issue to some changes between 6.6.0-RC5 and 6.6.0. both versions use spring boot 2.7. so its not a boot problem. the changes in the cas code cause config files from classpath:config not being loaded even when are specified in src/main/resources/application.yml spring.config.additional-location=config/
can someone advice me how to force cas auth again when
doChangePasswordparameter is present? the problem is when a user is not logged indoChangePasswordtriggers password change flow after successful login. but if a customer is already logged in (TGC) it doesn't show login view in order o change password but redirects directly to the service
@mmoayyed can you please advice, how to trigger password change when sso is active (valid TGC present)? I tried renew=true but it just prompts for password and after login it redirects back to the service=foo instead of showing password change flow. so the only current way is to logout customer and redirect him to cas change his password. but if he cancels the password change action then he kind of logged out for nothing and has to log in back. not the best UX I guess. so the best would be i guess, to redirect to cas with renew=true&doChangePassword, after successful login, perform change password operation and logout (quite common flow on many portals)
1 reply
Hi, we want to use apareo cas as oidc server , but we have the problem, that attribute in the scope email, openid are returned as array not as string. we use ldap as datasource
cas.authn.attribute-repository.ldap[0].attributes.mail=email
cas.authn.attribute-repository.ldap[0].attributes.cn=name
cas.authn.attribute-repository.ldap[0].attributes.sn=family_name
cas.authn.attribute-repository.ldap[0].attributes.cn=name
cas.authn.attribute-repository.ldap[0].attributes.givenName=given_name
in the logs we see
attributes={name=[customuser2], given_name=[Elke], family_name=[Tretsuser], email=[foo2@bar.com]}, id=customuser2, scopes=[email, openid, profile], client_id=clientid}
cas version 6.6.0
org.apereo.cas.pm.web.flow.actions.PasswordChangeAction and may be other actions are not localized which leads to such inconsistent ux. any suggestions on how to link those with messageSource?
DefaultCommunicationsManager uses messageSource to localize email subject and message property resolution using #{foo.bar} works but the locale is wrong. the email body is correct because of the EmailMessageBodyBuilder uses resolved locale but communication manager never receives locale to use, thus, it uses default locale. so the email is body is in german but subject in english. public class DefaultCommunicationsManager { ....
return messageSource.getMessage(matcher.group(1), args.toArray(),
"Email Subject", ObjectUtils.defaultIfNull(emailRequest.getLocale(), Locale.getDefault()));
...
}public class SendPasswordResetInstructionsAction { .....
val emailRequest = EmailMessageRequest.builder().emailProperties(reset)
.principal(person)
.to(List.of(to)).body(text).build();
return this.communicationsManager.email(emailRequest);
}
There is a new vulnerability with a very high CVE score (9.8) reported for Apache Commons Text. The vulnerable version is anything, between v1.5 till 1.9. In the current (latest) Apareo CAS 6.6 the version of the library which is in use is 1.9. Is there a plan for this to be upgraded to v1.10 or later?
https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2022-42889
Looks like they updated it, just haven't made a release
2 replies
2 replies
def run(Object[] args) {
def ALLOWED_EMAIL_DOMAIN="whatever.com"
def attributes = args[0];
def id = args[1];
def service = args[2];
def logger = args[3];
def suffix = "@" + ALLOWED_EMAIL_DOMAIN.toLowerCase()
if (id.toLowerCase().endsWith(suffix)) {
return id.substring(0, id.length() - suffix.length())
}
return id}
Hello All, I am trying to build cas management app and when I enable AWS service registry like dynamodb or S3, I get the following error, for dynamodb it creates the table and then fails to start:
APPLICATION FAILED TO START
Description:
Parameter 1 of method restAuthenticationService in org.apereo.cas.rest.config.CasCoreRestConfiguration$CasCoreRestAuthenticationConfiguration required a bean of type 'org.apereo.cas.authentication.MultifactorAuthenticationTriggerSelectionStrategy' that could not be found.
The injection point has the following annotations:
- @org.springframework.beans.factory.annotation.Qualifier(value="defaultMultifactorTriggerSelectionStrategy")Action:
Consider defining a bean of type 'org.apereo.cas.authentication.MultifactorAuthenticationTriggerSelectionStrategy' in your configuration.
Anyone have any ideas why this would be happening? if I remove the cas management AWS dependency it works fine.
Hi Guys! I was perplexed by SLO for a long time.
I read the doc,I understand when CAS logout, the CAS server will callback the CAS Client's URL to remind the Clinet to destroy the session.And the Client should definition an API to destroy the local session.
I followed that , but the Client never received the callback.
I want to know,if the CAS server will callback the Client?
I want to know,if the CAS server will callback the Client?
It should. Check the nginx/apache log, you should see something like:
172.20.0.83 - - [16/Nov/2022:02:37:02 +0100] "POST /EsupUserApps/login?target=https%3A%2F%2Fent.univ-paris1.fr%2Faccueil%2F HTTP/1.1" 200 0 "-" "Apache-HttpClient/4.5.6 (Java/11.0.16)"
hi,
Probably a neophyte question but I can't find the simple way to define an attribute from a header field with "trusted" authentication.
CAS authentication is simply done through an apache reverse proxy which defines an "APP_REMOTE_USER" header that I easily retrieve via:
cas.authn.trusted.remote-principal-header=app_remote_userThe proxy also provisions a 'GRP_FROM_APP' business header which I can clearly see in the debug log :
2022-11-17 10:02:56,702 DEBUG [org.apereo.cas.adaptors.trusted.web.flow.PrincipalFromRequestHeaderNonInteractiveCredentialsAction] - <Available request headers are [{referer=[https://XXXX], x-forwarded-port=[443], app_remote_user=[T01722], host=[XXX], connection=[Keep-Alive], cache-control=max-age=0],
....
grp_from_app=[017,223], cookie=[JSESSIONID=9Pt2j2GDjZTpmG09pVb525jCRRgVtTznCW8kgWDvJM8H2GJfGSMn!...
Chrome/107.0.0.0 Safari/537.36], sec-fetch-dest=[document]}]. Locating first header value for [app_remote_user]>
2022-11-17 10:02:56,705 DEBUG [org.apereo.cas.adaptors.trusted.web.flow.PrincipalFromRequestHeaderNonInteractiveCredentialsAction] - <Remote user [T01722] found in [app_remote_user] header>
2022-11-17 10:02:56,706 DEBUG [org.apereo.cas.adaptors.trusted.web.flow.BasePrincipalFromNonInteractiveCredentialsAction] - <User [T01722] found in request>
2022-11-17 10:02:56,708 DEBUG [org.apereo.cas.adaptors.trusted.web.flow.BasePrincipalFromNonInteractiveCredentialsAction] - <Attributes [{}] found in request>But there, I can't recover this header to transform it into an attribute. It looks like "Trusted" authentication doesn't provision any attributes other than the principal.
What method would allow me to add this attribute to cas clients?
Thanks
I see you are working on Trusted attributes and thank you very much for that.
I didn't understand what the ShibbolethExtractor class was doing here and I understand better now that you removed it (probably an "old" reason).
However, in your new implementation, I don't understand why relying on a "PREFIXheader" (AJP) => "Prefix_Value" model for the extraction, à la Shibboleth, which doesn't seem to me to be the most common.
Why not have simply, or at least also, a simple extraction based on the exact names (or regex) of the headers (a java.util.List "header1, header2, header3...") with the same logic as the releaseattribute parameters?
This method seems to me more suitable for authentication via trusted.
Regards
Hello,
I upgraded to the latest version (7.0.0-snapshot) and now I easily and directly retrieve ALL the attributes present in the Trusted header.
I admit that I didn't easily understand how to use attribute-header-patterns in cas.properties.
In my case, I used:
cas.authn.trusted.attribute-header-patterns=(APP_REMOTE_USER|GRP_FROM_APP)->(.+)Not very intuitive but is it correct ?
Regards
Hi all,
Could someone please assist with the below issue I am facing? Thank you.
Here is my environment (note that the frontend and backend are served from different domains):
- CAS Server (implementing CAS 2)
- Frontend
- Backend
Tech stack:
org.jasig.cas.client:cas-client-core v3.4.1
spring boot 1.5.4
Java 1.8
Here is what is happening:
(1) user accesses the frontend, but there is no session cookie yet, so the user is redirected to CAS login page served by the CAS Server.
(2) user enters correct login, and the CAS login page redirects to the ticket validating endpoint (login/cas?ticket=XXX....) served by the backend.
(3) the backend responds with a session cookie and redirects to the frontend URL.
The problem is that the cookie in step 2 is only valid for the backend domain, and not the frontend domain so all the requests from the frontend fail because the cookie isn't sent to the backend in requests. Anyone know a way around this? Can the cookie somehow be made cross-domain? Perhaps via configuration on the CAS Server? Please excuse my lack of knowledge w.r.t. the CAS protocol. Thank you
Is there a way to run CAS, in HA mode, without using sticky sessions, for OAuth/OIDC protocols (as mentioned in this comment: https://groups.google.com/a/apereo.org/g/cas-user/c/mzO_JpUqU3A/m/sMWhnIKrEgAJ)?
All CAS instances are already using a shared Ticket Storage, but, it seems that the Authorization Code Flow is not working properly without session affinity feature enabled in LB.
Thank you!