IPv6 SLAAC bricks keycloak

Apps & App Center
1

I have a functional UCS and Keycloak VM running.

If I add SLAAC to my IP configuration, it kills Keycloak. Trying to browse to Keycloak gives a proxy error.
image
image

So docker/keycloak or something doesn’t like IPv6. Okay fine. So I turn off SLAAC. Turns out Keycloak is still bricked at this point.

I’ve repeated these steps to confirm it is consistent, restarting the entire server between each step.

Looking at the keycloak docker container log, looks like it is barfing trying to resolve the FQDN of UCS in order to connect to postgres.

2024-02-24 16:26:43,746 ERROR [org.keycloak.quarkus.runtime.cli.ExecutionExceptionHandler] (main) ERROR: Property assignment of external_addr in TCP with original property value (FQDN redacted) and converted to null could not be assigned
2024-02-24 16:26:43,746 ERROR [org.keycloak.quarkus.runtime.cli.ExecutionExceptionHandler] (main) ERROR: Conversion of external_addr in TCP with property value (FQDN redacted) failed

Recovery right now is via VM snapshot. Anyone know how I could recover from this via CLI?

2

I have the exact same behaviour. I did the migration from UCS Server 5.0-9 err1212 to 5.2-0 err22 in small steps. My system has an IPv4 & IPv6 address. After the last step of the migration “Upgrade password encryption from md5 to sram-sha-256” the the Keycloak Admininterface can’t be access anymore. I did the same migration with a clone of my UCS Server with only an IPv4 address > The problem doesn’t occur.

Removing the IPv6 address and restrart the server doesn’t fix the problem.