EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  <20222023  2024  2025  Index 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  <20222023  2024  2025 
<== Date ==> <== Thread ==>

Subject: Re: 'CAS-client' failed in ../db/dbEvent.c line 540
From: Michael Davidsaver via Tech-talk <tech-talk at aps.anl.gov>
To: Oleg Makarov <makarov at anl.gov>
Cc: "tech-talk at aps.anl.gov" <tech-talk at aps.anl.gov>, Ralph Lange <Ralph.Lange at bessy.de>
Date: Wed, 7 Sep 2022 12:53:57 -0700
On 9/6/22 15:57, Oleg Makarov wrote:
Michael, see my answers below.

On 9/6/2022 4:48 PM, Michael Davidsaver wrote:
On 9/6/22 13:19, Makarov, Oleg A. via Tech-talk wrote:
Hi Jeffrey and Ralph,


I am observing occasional/intermittent error in the vxWorks EPICS IOC. Error occurs during connection from CA client to the PVs in the IOC.
mvme6100 IOC is running on EPICS version 7.0.6.1. We have CA clients using Java JCA library 2.4.1 and pyepics 3.5.0.

Also, standard question.

What drivers and/or support modules are loaded into this IOC?


Does this IOC have an access security file (ACF) loaded?
*No.*

Next time this happens, please try to collect a stack trace.  With this additional
information, we may be able to distinguish between the other two call sites.

While this crash appears differently than lp:541371 (link below), my suspicion is
the same.  Some kind of race condition in the CA server (RSRV) on client disconnect.

So you might be able to provoke the crash with a lot of very short lived clients.
(eg. maybe some script is crashing in a tight loop?)


If so, does are any dynamic rules used?


Does it have any external CA links, connected or otherwise?
*Yes.*

Never mind about this.  I was confused.  I don't think CA links from this IOC
have anything to do with this issue.


dbcar "" 2
*see attached "dbcar.txt" file*

A call to 'assert(status == epicsMutexLockOK)'
     by thread 'CAS-client' failed in ../db/dbEvent.c line 540.
EPICS Release EPICS R7.0.6.1.

fyi.

https://github.com/epics-base/epics-base/blob/abfe30ef13ad60b43ae6f257994c8e7a503d4f3a/modules/database/src/ioc/db/dbEvent.c#L540

This is a different location than:

https://bugs.launchpad.net/epics-base/+bug/541371


References:
'CAS-client' failed in ../db/dbEvent.c line 540 Makarov, Oleg A. via Tech-talk
Re: 'CAS-client' failed in ../db/dbEvent.c line 540 Michael Davidsaver via Tech-talk
Re: 'CAS-client' failed in ../db/dbEvent.c line 540 Oleg Makarov via Tech-talk

Navigate by Date:
Prev: Re: Basic iocStats database Mark Rivers via Tech-talk
Next: Understanding EPICS_CA_ADDR_LIST and caget Sean Leavey via Tech-talk
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  <20222023  2024  2025 
Navigate by Thread:
Prev: Re: 'CAS-client' failed in ../db/dbEvent.c line 540 Oleg Makarov via Tech-talk
Next: Understanding EPICS_CA_ADDR_LIST and caget Sean Leavey via Tech-talk
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  <20222023  2024  2025 
ANJ, 14 Sep 2022 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions ·
· Download · Search · IRMIS · Talk · Documents · Links · Licensing ·