New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Cannot start using "start --optimized" after upgrading to a newer version #29106
Comments
Your trace looks a bit as if there is a second Keycloak container it is trying to form a cluster with but this does not work. Could it be that the upgrade tries to put a new Keycloak container next to the existing one to do the update? This is generally not supported for with different Keycloak versions. To update, you first have to stop the old container and then start a new one. |
@sschu I noticed that too but I don't know where the 2nd entry is coming from. I stop the stack, update the keycloak container and start it back up. There aren't two servers online at once. It always shows two, never 3 or more, no matter how often I restart or change versions though. The weird thing is that this error doesn't appear when using |
start-dev creates only a local cache. |
Ah okay, that explains it. Is there a way to clear the cache? Is that even a good idea? I still don't understand why it only happens when I change the version. Even if I completely kill and delete the container (retain the database), as long as I start it up with the same version there's no problem. It probably doesn't like migrating from an older version if there's unreachable hosts configured in the cache. |
@madmurl0c the exception your are showing should only happen if there is another member of the cluster running with a different version. a0c08860e31b-49509 is the local address, then you can see another one:
|
I feel kind of stupid now for not thinking of this earlier but I think I know where the 2nd instance comes from. I use traefik as a reverse proxy and have the I know that I should probably have the test-system running on different hardware but is there a way to separate the caches? Give them a name or prefix maybe? Something like Thanks for pointing me in the right direction @shawkins |
The operator uses -Djgroups.dns.query and a headless discovery service to control this - as pointed out here #21067 (comment) - you can create separate services per version to segregate the members. I'm not sure what's the simpliest way to do this otherwise - cc @mhajas |
@madmurl0c you're mixing different Infinispan versions in the same cluster and that is not supported by Infinispan. A wire format of one of the messages was changed to fix an issue and the old version is unable to parse it. That's why you have |
Thanks everyone for your input. I worked around this issue by adding |
Before reporting an issue
Area
infinispan
Describe the bug
Hey everyone, I'm experiencing an issue trying to upgrade keycloak.
I've been using version 21.1.0 for quite a while but I cannot upgrade to any newer version.
I narrowed it down and noticed that it does start when I use "start-dev" instead of "start --optimized".
I build my optimized container like this:
Version
21.1.0, 24.0.3
Regression
Expected behavior
Keycloak should start up
Actual behavior
Following error message:
How to Reproduce?
I created a new docker-compose stack like this:
Then try to upgrade like this:
Switching back to the older version and it starts up just fine.
Anything else?
No response
If you need any more info just let me know. Thanks for your help :)
The text was updated successfully, but these errors were encountered: