Experimental Physics and
Industrial Control System

"Jeff Hill" <[email protected]> · Thu, 10 Apr 2008 09:00:19 -0600

> But as Rodney mentioned it, this is also true for the same base 3.14.9
> built on a different linux versions.
> Ex:	
> caget from 3.14.9 on RedHat 8.0 segment faults for waveforms.
> caget 3.14.6 on RH8 is ok
> caget from 3.14.9 on fedora core 6 is ok.

This is the first that I have heard about this. Unfortunately we have only
RHEL here. We use R3.14.8.2 in the LANSCE control room with recent Red Hat
Enterprise Linux versions without issues. We however don't use medm here (we
use edd/dm and are considering switching to edm).

Could you attach the debugger to a debug build and send the stack trace when
it seg faults?

Jeff

> -----Original Message-----
> From: [email protected] [mailto:[email protected]]
> On Behalf Of Emmanuel Mayssat
> Sent: Wednesday, April 09, 2008 3:47 PM
> To: Rodney R. Porter
> Cc: [email protected]
> Subject: Re: MEDM CA Exception Limit
> 
> Rodney and others,
> 
> I have noticed that the CA behave differently (work, return
> errors/warnings, or doesn't work at all) when the epics-base is changed
> between 3.14.6 and 3.14.9
> 
> But as Rodney mentioned it, this is also true for the same base 3.14.9
> built on a different linux versions.
> Ex:	
> caget from 3.14.9 on RedHat 8.0 segment faults for waveforms.
> caget 3.14.6 on RH8 is ok
> caget from 3.14.9 on fedora core 6 is ok.
> 
> I suspect this issues is related to those I am experiencing with medm
> and edm.
> 
> Regards,
> 
> --
> Emmanuel
> 
> 
> 
> On Thu, 2008-03-27 at 09:01 -0700, Emmanuel Mayssat wrote:
> > You may want to look at the -noMsg option of medm.
> > Are you sure the problem you are having is related to the version OS?
> >
> > I have several problems with medm (and edm BTW), but in my case it seems
> > to be related to the version of the epics base I am building against.
> > My medm (3.1.1+) works when build on epics base 3.4.6 and but not with
> > 3.4.9 ( compile, execute, but never connect to IOC)
> > Conversely, my edm (1_10_1d+) works when compiled against 3.4.9 but not
> > with 3.4.6 (doesn't compile easily)
> >
> > Good luck,
> > --
> > E
> >
> >
> >
> > On Wed, 2008-03-26 at 17:32 -0500, Rodney R. Porter wrote:
> > > I get the following error from MEDM on Fedora Core 4, with similar
> > > unexplained MEDM communication hangs on Fedora Core 6:
> > >
> > > medmCAExceptionHandlerCb:  Channel Access Exception:
> > > Too many exceptions [25]
> > > No more will be handled
> > > Please fix the problem and restart MEDM
> > >
> > > As far as I have been able to tell, the exceptions are generated when
an
> ioc
> > > is restarted.  Does anyone have any other experience with this error?
> > >
> > > Rodney Porter
> > > Nanoprobe Beamline
> > >
> >

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System