EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  <20062007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  Index 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  <20062007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: Re: IOC lost all comms
From: Andrew Johnson <[email protected]>
To: Al Honey <[email protected]>
Cc: [email protected]
Date: Wed, 26 Apr 2006 16:17:51 -0500
Al Honey wrote:
One of our IOCs became unresponsive to CA requests.

aux2:1041 Tue Apr 25 19:12:02 2006 filename="../taskwd.c" line number=159
aux2:1041 Tue Apr 25 19:12:02 2006 task 3718270 cbLow suspended                    ß
aux2:1041 Tue Apr 25 19:12:35 2006 filename="../dbNotify.c" line number=257
aux2:1041 Tue Apr 25 19:12:35 2006 S_errno_EINTR dbNotifyCancel had semTake timeout
aux2:1041 Tue Apr 25 19:12:37 2006 filename="../dbNotify.c" line number=257
aux2:1041 Tue Apr 25 19:12:37 2006 S_errno_EINTR dbNotifyCancel had semTake timeout
aux2:1041 Tue Apr 25 19:12:38 2006 filename="../dbNotify.c" line number=257
aux2:1041 Tue Apr 25 19:12:38 2006 S_errno_EINTR dbNotifyCancel had semTake timeout
aux2:1041 Tue Apr 25 19:12:42 2006 0xlename="../dbNotify.c" line number=257

where the semTake timeout repeated until the IOC was rebooted.

Unfortunately those that were party to the problem did not attempt to diagnose, so there is no other feedback.

Does anyone know under what conditions the cbLow task will become suspended?

A task suspend often means there was a bus error, or some that other processor level exception occurred which vxWorks can't fix itself. The cbLow task will be used at record I/O completion time by almost every asynchronous device support where the record's PRIO field is set to Low. The callback tasks are also used by SEQ records for processing the input/output link pairs, by bo records that specify a positive value for HIGH, by calcout records that set ODLY and by histogram records (not quite sure what for though).


HTH,

- Andrew
--
There is no S in exprexxo.

References:
IOC lost all comms Al Honey

Navigate by Date:
Prev: RE: General lock question Jeff Hill
Next: RE: StripTool bogs down host? Chu, Paul C.
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  <20062007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: IOC lost all comms Al Honey
Next: New Device Support for Various Items John Dobbins
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  <20062007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 02 Sep 2010 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·