[keycloak-user] Issue with Keycloak startup in AWS as a docker

[Jagannadha Rekala]

Hello,


We are running Keycloak (database: Oracle 12c) on AWS as a docker in EC2 instance. Keycloak deployment is successful via the cloud formation but the startup failed with Keycloak receiving the TERM signal. We have amazon-agent along with Keycloak's docker to spin a new container of Keycloak. Amazon-agent agent starts a new Keycloak container within a minute. The second time startup of Keycloak fails as well with a different error (where is is not able to create a new table while the name already exists - ORA-00955).


This is happening in random. Sometimes the Keycloak instance at the first time starts well, without any issues. I have attached the logs of Keycloak container starting first time after deployment and the second time the amazon-agent spins up a new container. Keep in mind that the logs are bottom-up, watch from the bottom to up for a sequence. Ultimately, from the logs we understood that it is trying to create the tables the second time but those tables already existing.


Please let us know if you have encounter this kind of issue or any pointers where the issue could be why the first time the container receives a TERM signal and why the second time it cannot overwrite the tables while migration strategy mentioned as update in the standalone.xml. Any help would be appreciated.


Thanks,

Jagan Rekala