Details
-
Bug
-
Resolution: Done
-
Major
-
1.1.0.Alpha11
-
None
Description
If you call the shutdown(reload=true) goal, followed by the deploy-artifact goal, it goes boom. The reload operation runs asynchronously, and temporarily takes down the management interface that the deploy-artifact goal needs to succeed, creating a race condition that has ended badly 100% of the time for me.
Why might a project need to reload then deploy? For instance, it might start the server, use execute-commands to set system properties required by an application, then reload and use deploy-artifact to deploy that application. Similarly, you might want to reconfigure SSL and then reload, as that change also requires it.
It's worth noting that this likely has a simple fix: just copy the "wait for server to be ready" code used in the start goal over to the shutdown goal's reload handling.