Bugzilla – Bug 7007
destruction/schedule issues
Last modified: 2010-04-30 09:18:59
You need to log in before you can comment on or make changes to this bug.
When client does not destroy but goes away (this also typically happen with raw workspace client), the shutdown routines in the scheduler have bugs that cause incorrect behavior. Below are log messages for search engines (to help troubleshooters): ERROR 1, this repeats often: Running time has expired, client requested default shutdown mechanism of Normal, setting target state to StagedOut ERROR 2, this repeats often: Problem sweeping resources: edu.emory.mathcs.backport.java.util.concurrent.FutureTask cannot be cast to edu.emory.mathcs.backport.java.util.concurrent.Callable
Fixed and committed for Nimbus 2.4 RC2 http://github.com/nimbusproject/nimbus/commit/e9f451c02cfa10177e4b8ab50aa29565433b7ea6
Original description is a little wrong: this only happened when there was a save-back request by default when the time runs out (not a terminate-by-default) and there was no client around to trigger destroy when the state of the VM moved to 'TransportReady' (i.e., unpropagated). This is why this regression was not seen in 2.3, it's not all that common of a situation.
