Scheduler Core Dump - Hyperthreading ?

Sometimes the scheduler crashes with core dump. After restart we find in the log, that a the last job in the log is listed [first with finished], then [with started] means the job order is nonnatural before crash.
This happes since migration to a new dual core system with windows server 2003. Maybe new semi-parallel features of the [CPU/OS] are the reason ? We discover the same error class with aged versions of MS Access where we see, that program execution is jumbled.

What do you mean by "nonnatural" order? What kind of jobs do you run? Have you tried to set all jobs to run detached?

Hello, thank you for that hint. Now all jobs are detached. We will post in this thread if the problem will persist.
To clarify the order of entries in the log: For a given job (start) is listed before (finished) as seen from the top of the list where the regular order should be that the (finished) entry should be most recent.

Otherwise Albert Einstein would be wrong, and you could be your own grandfather - booom - vacuum - no 24x7 scheduler ever invented.

A few days ago I posted a request reporting irregular crashes of 24x7 on Windows 2003. We found, that using remote desktop together with leaving the GUI of 24x7 open is the cause. Whenever you minimize to icon tray before detaching the system remains stable. Otherwise it will crash a few times a day.

Thanks for the status update.

By the way, you can run 24x7 as a service and user either 24x7 Remote Control or 24x7 Web-based Interface to manage and monitor jobs remotely. Remote Desktop will not be required in such case.