Re: Alarm stuck in "Async in progress" even after record processing is fine

[Marco Filho via Tech-talk <tech-talk@aps.anl.gov>]

Hi, Andrew, Jure

Thanks for the explanation, I will test with latest EPICS versions and see if it works.
Searching for DBR_PUT_ACKT I now managed to find some documentation about it also 🙂

Best regards,

Marco

---

From: Johnson, Andrew N. <anj@anl.gov>
Sent: Wednesday, November 26, 2025 19:55
To: Marco Filho <marco.filho@ess.eu>; EPICS Tech-Talk <tech-talk@aps.anl.gov>
Subject: Re: Alarm stuck in "Async in progress" even after record processing is fine

 

Hi Marco,

Acknowledging an alarm with the IOC won’t clear the AMSG field. As Jure wrote, there was a bug in the handling of that field which has been fixed in EPICS 7.0.9.

If you’re still curious about the ACKT field, there are three type codes DBR_PUT_ACKT, DBR_PUT_ACKS and DBR_STSACK_STRING plus associated data types provided by Channel Access and the CA client library to globally acknowledge alarms inside the IOC. Unfortunately the PVA protocol doesn’t implement anything like these though, so if you’re using PVA (since you also asked about AMSG) you can’t store your alarm acknowledgement states in your IOCs. The CA operations were used by the old Motif-based Alarm Handler client, but the newer alarm systems implemented for CSS/Phoebus keep their own alarm acknowledgement database. I don’t think they can use the CA acknowledgement mechanism at all.

HTH,

- Andrew

-- 

Complexity comes for free, Simplicity you have to work for.

On 11/21/25, 9:19 AM, "Tech-talk" <tech-talk-bounces@aps.anl.gov> wrote:

Hi, all.

I'm not sure if this is a bug or a feature: I have an IOC that hangs the record processing when the hardware disconnects. This causes the record to process over itself more than ten times, generating an "Async in progress" alarm message.

When the hardware finally reconnects, the record processing gets back to normal again, but the alarm message never disappears. AMSG and NAMSG keep displaying "Async in progress". ACKT is YES and ACKS is INVALID, although SEVR is already in NO_ALARM.

Shouldn't AMSG go to an empty string? Is there a proper way to acknowledge this alarm?

I found in the documentation:

"The ACKT field specifies if it is necessary to acknowledge transient alarms."

But I couldn't find how is this acknowledgment made.

Thanks for any help,

Marco