Re: Disable alarm at IOC via PVAccess

[Michael Davidsaver via Tech-talk <tech-talk@aps.anl.gov>]

Hi David,

I'm wondering if there is a jargon mismatch here.  When you say "disable an alarm", do you mean in sense that "this condition should never result in an alarm"?  Or instead "Operations has evaluated the situation, taken action, and would like to focus on something else now thank you very much"?

On 11/20/25 1:47 PM, David Bracey via Tech-talk wrote:

A PV really has four distinct alarms – all four severities would need to be set NO_ALARM to really “disable alarms” for the PV

Oh, many many more conditions than this!  Off hand there are calculation alarms, undefined value (UDF), and scan disable.  Search through the epics-base source for "recGblSetSevr".  And there are even more driver specific conditions.

To re-enable alarms, you would have needed to have remembered the prior value to restore it (or just YOU’LL GET MAJOR AND LIKE IT)

It sounds to me like you don't want to do this :)

I suspect that what you want would in the epics lingo is to "acknowledge" an alarm (the second definition above).  Originally this was done in the IOC, and still can be.  However, these days it is typically handled in an alarm manager service.  cf. the Phoebus Alarm Server (or its wonderfully named predecessor BEAST).

https://control-system-studio.readthedocs.io/en/latest/services/alarm-server/doc/index.html