Experimental Physics and
Industrial Control System

"Chestnut, Ronald P." <[email protected]> · Mon, 14 Sep 2009 09:52:21 -0700

At SLAC we have been using IRMIS heavily, and depend on it for up-to-the-minute PV locations. The IRMIS database is also useful for noting items which stray from the naming convention, or may have a typo.

It is also useful to verify archiver or alarm PV lists.

Ron Chestnut 

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of quock
Sent: Sunday, September 13, 2009 6:49 AM
To: Noboru Yamamoto
Cc: epics Techtalk
Subject: Re: list all process variables

Hi Noboru,

The beauty of storing information about EPICS PVs (along with IOC 
architecture, installed controls hardware, etc.) in a relational 
database is that it allows one to easily track the history of an IOC's 
configuration such as when new PVs were added to an IOC and when PVs 
were deleted from an IOC.

EPICS-IRMIS workshops held over the past few years have drawn much 
interest, support and collaboration from several accelerator facilities 
(CLS, SLAC, BNL, MSU-NSCL and more). An IRMIS workshop was held this 
past spring at Brookhaven National Laboratory and was led by Gabriel 
Carcassi. Gabriel and I will both be attending the ICALEPCS 2009 
conference, perhaps Gabriele and I can arrange a side-session meeting or 
an evening meeting to discuss future enhancements for IRMIS and recent 
updates to IRMIS version 3 sometime during the week of the ICALEPCS 
conference.

Debby Quock
Controls Group
Advanced Photon Source
Argonne National Laboratory

Noboru Yamamoto wrote:
> Hi,
>
> Patrick Thomas wrote:
> > Hi,
> >
> > I was wondering if there was anyway to get a list of all the process
> > variables currently loaded in all the IOCs. I read that when a request
> > for a particular process variable is made, each IOC gets the request 
> and
> > has to look in its databases to see if it contains it. Is it 
> possible to
> > instead have it return all the process variables it does have?
> >
> > I was also wondering if anyone has had trouble getting the sound to 
> work
> > when compiling the alarm handler on Linux?
> >
> > Thank you,
> > Patrick
>
> A thread in TechTalk titled "Listing channels" by David Dudley dated 
> May 06, 2008 and responces  may also help you.
>
> In these responces, I reporeted my device support for waveform record:
> devWfDbRecfection
>
>    1. A device support for waveform record. It get a list of record 
> types/fields of specified record/name of reocrds from EPICS DB using 
> dbStaticLib.
>    2.Now Supports CHAR and UCHAR as FTVL. It returns comma separated 
> list of informaiont as char array.(2009.3.20)
> and tar.gz file is available at:
>   http://www-acc.kek.jp/EPICS_Gr/products/Records/DbReflectionApp.tgz
> To achieve your goal, you need to install records using this device 
> support in ALL IOCs you have.
>
> Alternatively,
> You can use dbDumpRecord and dbDumpRecorType commands,or just "dbl", 
> in iocsh to dump those information to a file. Again you need to add 
> these command(s) in every st.cmds in your system.
>
> If you can safely assume you know the location of all start-up command 
> files of running IOC, a PV Crawler can find all records in your system.
> In this case, we also assume that database files used in the start up
> commands have not changed since last reboot of IOCSH.
>
> Anyway, It is interesting  that this kind of question raised several 
> times in TechTalk, and several different solutions were shown. Doesn't 
> it means we need to do something in the Collaboration, "provide FAQ on 
> this subject" or "Define new protocol
> to retrieve these information" for examples?
>
> Regards,
>
> Noboru Yamamoto
> J-PARC control group/J-PARC center(KEK & JAEA)
> EPICS group/KEK
> JAPAN
>
>
>

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System