Best practice in these situations is, to first set the client to not fetch work and abort the jobs not started and or not completed. That would signal the scheduler to reissue immediately.

Normally a client reset would retrieve all none aborted work, provided it's the same install. In this situation, they will eventually expire as 'No Reply' and be send to another cruncher at that time.

Techs really don't have the time to manually annul tasks, particularly now, where a project launch is feverishly worked on. Most probably this happens so often in a day that it would become a day task to trace them down, so thanks for letting us know but don't worry about it too much.

cheers