Repository navigation
Replies: 2 comments 6 replies
|
I've converted this to a discussion as this almost certainly a configu
8000
ration issue. For example, the |
|
Are you able to reproduce this issue locally using a minimal configuration? Essentially, if you can reproduce this locally try to strip away options until it starts working, or start clean and add your options until it breaks to find what the cause of it is. If you cannot reproduce this locally there must be caused by some kind of restriction of your cloud environment. If I had to guess I think the path |
|
Thanks for your response, @jonkoops! You pointed me in the right direction and I was finally able to resolve this issue. A closer examination of the browser logs revealed that the full path of the file name was getting corrupted with some additional non-printable characters. I think this was because I was attempting to launch Keycloak using the ALB domain name which is a very long computer generated name with lots of odd characters. I created a simple DNS CNAME for the ALB and was able to successfully launch the console using it. Although the ALB domain name is valid, there may be an issue in the way that long, complex domain names are handled by the Keycloak application. |
|
@razrdog Hi. Could you provide a bit more details about additional non-printable characters please? How did you find them? We have exactly the same problem in our Keycloak after upgrading from Keycloak 24 to Keycloak 26. TLS is terminated at ingress level. We see no errors anywhere, last request goes to |
|
@gim- are you using a domain name that matches your TLS certificate? In my case, I was attempting to test 8000 using the ALB domain name which doesn't match my TLS certificate. This was causing Keycloak to get stuck loading with the 204 status. It started working after I created a DNS CNAME record for Keycloak that matched my TLS certificate. |
|
@razrdog I managed to resolve this, it was not related to TLS certificates or anything like that in our case. We have a custom extension of the default hostname provider in Keycloak which resolves hostname from request headers (if present) to support multiple hostnames, and it was taking the port number 443 from keycloak/services/src/main/java/org/keycloak/url/HostnameV2Provider.java Lines 152 to 155 in e4987f1 And it unintentionally bypassed the port sanitaze part in the This resulted in the following output in the
The mismatch between frontend URL and Admin URL was causing the same symptoms. Getting rid of "443" port in the frontend URL fixed it. |
|
I am having this same issue, with the following output from the debug page
Ive tried:
The login flow for SOLVED!This was the issue, and its a clever lil gotcha, the problem was here:
Note the caps vs lowercase, turns out its case sensitive. |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Before reporting an issue
Area
admin/ui
Describe the bug
The keycloak console stops loading and just shows the continuous rotating circle. As shown below, it returns HTTP status code 204 when attempting to retrieve /realms/master/protocol/openid-connect/login-status-iframe.html/init.
Keycloak is running as an AWS ECS Docker container behind an Application Load Balancer which provides TLS termination as a reverse proxy. Environment variables are set to the following:
Version
26.0.8
Regression
Expected behavior
The console should load successfully.
Actual behavior
The console is stuck attempting to retrieve /realms/master/protocol/openid-connect/login-status-iframe.html/init with an HTTP status code 204 (No Content). No errors are generated.
How to Reproduce?
I've attempted to change the hostname variables with no success.
Anything else?
No response
All reactions