We are in the process of developing an audit cycle for our DW. In the event that we are out of balance with the source system we want to turn off the WF scheduler and delay the reports from running until the problem has been corrected.

Is there any way to do this automatically or does it mean manually turning off the scheduler and then turning it back on when the problem is resolved?

Michelle, here's what we have done.
in each schedule there is a slot for PREPROCESSING fex to be named, actually 2 slots,
(and a couple of slots for post processing fexes to be run).
1st.For each job, we have a single program named in that preprocessing fex #1 slot that can set the value of &&KILL_RPC to 'Y' if we need to kill that job. We call that program KILLALL.fex
and it just has a comment in it
-*SET &&KILL_RPC = 'Y' ;
We can uncomment that line and invoke the kill if we need to. We can also make that kill date specific...set the value to Y if within a certain date or time range...whatever we need. THis was useful over the 4th of July when we were closed..i just set the variable to Y or N based on the &YYMD being >= july 5th; So we were all quiet over the holiday, and i didn't have to touch anything.
2nd...on that second PREPROCESSING fex line, we have a program that is job-specific, and can test stuff...eg ,run to see if updates have been made... and then SET &KILL_RPC = 'N' if everything is fine, or 'Y' if we gotta kill it.
Note: These PRE-fexes are TASK specfic...so you gotta set them up for each TASK in a SCHEDULE.
For us, this works like a charm. There are times when we have to kill all the ETL jobs and times when we have to kill only the emails.
Its very flexible this way.
Hope that helps.

note: &&KILL_RPC is the system variable that kills a caster job...its not something i made up

Hello Michele

We have developed a technique that permits us to do the following:

In the preprocessing fex we check for the presence of daily data (it is often late). If the data is late then we delay the schedule by 10 minutes and try again. User gets an email saying "sorry data is late... ". User only gets this email once. Caster will keep trying every ten minutes. Once the data is there report is distributed and the schedule is set for the correct time the next day.

All of this is totally automatic. It involves updating the NEXTRUNTIME field in BOTSCHED. Let me know if you want more details....

Jodye

Thank you everyone! Susannah - your response is intriguing and I will give it a shot. Jodye - can you give me more details? This method appears to be a little more straight forward.

I am also weak on the GUI - I have over 23 years experience in core Focus but the GUI drives me nuts! So, hopefully I can figure this out without too much loss of hair.

Hello Michele

I know what you mean about the GUI. Long live textpad!

I have already posted our technique for manipulating the nextruntime in botsched.

Look at this post...

https://forums.informationbuilders.com/eve/forums/a/tpc/f/7971057331/m/
8461028331/r/8461028331#8461028331

Scroll down to my very long post...

Take a look at that. If you have any questions please drop me a line...

We have hundreds of jobs every day that are handled by this process and it seems to be working fine. But like I said, IBI does not like us to manually play with the NEXTRUNTIME field in Botsched...

Hope this helps...

Jodye

This message has been edited. Last edited by: Kerry,

Thanks Jodye - that was very interesting reading and we will play with it this week. My staff is kind of excited about it. I hadn't attended the session that you and your boss were presenting at Summit but wish I would have to connect a face with your profile.

Thanks again!

Hi Michele

I hope it works for you. By the way, we no longer use an sql trigger. We simply calculate the current nextruntime in the prefex...

Good luck.

Jodye