Experimental Physics and
Industrial Control System

"Davidsaver, Michael" <[email protected]> · Wed, 8 Jun 2011 11:14:53 -0400

Hi John,

I've run into this problem before.  In my case I could drop events if
the rate is too high.
So I added a counter for each CALLBACK which is incremented when it is
requested, and decremented
When it completes.  If the counter is too high then callbackRequest()
isn't called.

Regards,
Michael

> -----Original Message-----
> From: [email protected] [mailto:tech-talk-
> [email protected]] On Behalf Of John Dobbins
> Sent: Wednesday, June 08, 2011 9:02 AM
> To: EPICS Tech-Talk
> Subject: general purpose callback tasks : callbackRequest()
> 
> Greetings,
> 
> I have an IOC on which all records stopped updating for 5 seconds at a
> time followed by resumption of normal operation. The IOC would operate
> normally for hours at a time and then would suffer a burst of stop/go
> activity lasting several minutes.
> 
> The IOC has multiple thread implementing asynchronous device support
> and
> uses the general purpose IOC callback tasks. I think the problem is
> that
> callbackRequest(), which can fail silently, was overflowing the
> callback
> queue.
> 
> My problem appears to be fixed by increasing the callback queue size.
> 
> My question: Is there a way to know about the callback queue status or
> otherwise avoid this situation?
> 
> John Dobbins
> 
> Cornell Laboratory for Accelerator-based Sciences and Education
> Cornell University
> Ithaca, NY

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System