webMethods Version: 8.2 (upgrading to 9.12 in January 2017)
We have been having a periodic issue where one of our Oracle databases (JDBC Adapter) will get restarted, and the one, simple-interval task on the Scheduler will get stuck processing somewhere in the flow of the associated service. The database comes back up, but the thread never resolves or times out. When we try to cancel and then kill the system thread, it just creates two new threads for those tasks, which never get resolved either. Trying to disable the associated adapter gives the same result, creating a new thread that never resolves.
Our only solution so far has been to stop and start the services on the server: Integration Server, Broker manager & server, and MWS.
Of course, we would ideally be able to pause the scheduler before these database restarts, but we’re still researching why the database is restarting at all. And besides that, we have other Oracle databases that can error out, restart, whatever - and the related scheduler tasks and services don’t get hung up.
Any thoughts on what we can do to help this? I’ve added time-out values to all of the services, but the note on time-out says it will throw an exception after the service step “completes”. I’d say that our problem is that the service step is never completing.