Q207049: Task scheduler corrupted on restart for 64 bits linux client machines – Raynet Support Panel

Symptoms

The scheduled tasks on 64 bits unix client machines suddenly stop running with no apparent reasons.

Cause

ndtask erase the scheduled events' triggers if the logon/restart trigger is not present for the events

Steps to Reproduce

- Make sure ndtask is running using "etc/init.d/ndtask start"
- Make sure the schedule is installed using "/opt/managesoft/bin/ndschedag /var/opt/managesoft/scheduler/schedules/sched.nds"
- Verify the schedule was installed properly using "/opt/managesoft/bin/ndschedag -e"
This should provide a list of events and their next run times as dates and times. For example "Update Client Settings 04/28/2012 04:36:00 PM"
- Restart ndtask using "/etc/init.d/ndtask restart"
- Verify the schedule again using "/opt/managesoft/bin/ndschedag -e"
You will probably see the next run times are all "On Connection" with a time offset. For example "Update Client Settings On Connection + 10 minute(s)".

This would confirm that the task scheduler, ndtask, is corrupting the task files after starting up.

This could also be validated by stopping ndtask (i.e. /etc/init.d/ndtask stop), making sure the schedule is installed and good (i.e. ndschedag .../sched.nds and ndschedag -e) and then running the ndschedag -o "Startup=True". When running ndschedag -e after this, you will see the time based triggers are missing.

Workaround

Make sure the logon/restart trigger is present for all the scheduled events.