[ https://issues.jboss.org/browse/WFLY-11860?page=com.atlassian.jira.plugin... ] Brian Stansberry commented on WFLY-11860: ----------------------------------------- You can do a reload but not a restart. The point of --read-only-server-config is the configuration you want is locked down, e.g. you have some 'approved' configuration that is controlled by whoever in your org does such approvals and that is what you are meant to use. You're not meant to try and persist changes and have them apply after restart, at least not without effort like starting pointing at a different config file. We do allow you to make config change to a server started with --read-only-server-config. Arguably we should not, but we do. The reason for that is it's reasonable to want to do something that affects config that is operational but not needing permanence, like changing a log level on a running server to get some diagnostic data. If you make other changes that are something you'd want persisted, then that's not an intended use of --read-only-server-config. (There's no practical way to discriminate between changes that are likely something you'd want persisted versus those that are not.) If you do make changes, we persist them in history. The point of that is to allow reload. Arguably we should not support that, but we do. Mostly we do because if you invoke an op that requires reload, the response will state that it requires reload. So then it seemed strange to not allow the reload. An alternative would have been to fail the op if it required reload. Adding a flag to enable that alternative behavior would be a reasonable RFE, as would be a setting to disallow any config changes at all. But those are RFEs as the current behavior is intentional. -- This message was sent by Atlassian Jira (v7.12.1#712002)