[ https://issues.redhat.com/browse/WFLY-14199?page=com.atlassian.jira.plugi... ] Brian Stansberry commented on WFLY-14199: ----------------------------------------- Thanks, [~jamezp]. This kind of thing seems like an argument for WFCORE-436. Although in general it seems like the basic WF design should allow for stop to be called in any case where some start was called on some type that has a start/stop kind of method pair. It might not be called when application deployment fails during boot, but when the server was reloaded any MSC services that had been started during that failed deployment attempt should have been stopped, allowing cleanup work to happen. And DeploymentUnitPhaseService itself will call undeploy() on any successfully executed DUPs if a subsequent one in its phase fails. Individual MSC services that fail during their own start() method may not be doing cleanup though. -- This message was sent by Atlassian Jira (v7.13.8#713008)