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

Subject: Re: CA communication problems
From: Christophe Moins <christophe.moins@ego-gw.it>
To: Jeff Hill <johill@lanl.gov>
Cc: tech-talk@aps.anl.gov
Date: Tue, 18 Jul 2006 15:25:49 +0200
Hello,

after deeper investigations, we have hopefully found out the origin of the problem.
When using the netstat command, we saw an anomaly at the level of the socket between the IOC and one of the archiver engines that were running.
The socket buffer was full on the archiver side, which prevented the ioc that remained stuck from sending other information.
We assumed that the problem could be linked to the size of the buffer that the archiver is using when it writes data to the disk.
We increased by a factor ten the size othe archiver buffer and since then the system is stable and we did not notice any more full socket buffer.


Thanks for your help
Christophe

Jeff Hill wrote:

I have some problems of communication between an IOC and a client, both
running on different machines on the same subnetwork on a similar
environment (Epics v 13.4.7 under linux redhat 9.0 kernel 2.4.20-8).

At the beginning of the startup of the IOC, process variable values can
be got without problems from a client (either a simple command line or a
php script), but after a while, the data is not available any more and
the following message shows up on the IOC side :

CAS: request from 192.168.101.21:55894 => "bad resource ID"
CAS: Request from 192.168.101.21:55894 => cmmd=12 cid=0xffffffff type=0
count=0 postsize=0
CAS: Request from 192.168.101.21:55894 =>  available=0x1        N=0
paddr=(nil)
CAS: forcing disconnect from 192.168.101.21:55894

Would someone have an idea of what's going on ?



I have no guesses at this point, but instead a few questions.


O Is this Linux running on a typical Intel x86 computer?

O What does the output from the "casr 10" command, typed into the IOC's
console when the failure is occurring, provide?


O I would also look at the cpu, memory, and socket consumption in the IOC's
host using top and netstat.

O Is 192.168.101.21 the IP address of the host for the client that isn't
able to connect, or does the host IP in the message vary?

O Does the IOC appear to be functioning correctly in other respects when
this occurs?

O Does the IOC have any layered products (such as drivers) installed when
this occurs?

Jeff



-----Original Message-----
From: Christophe Moins [mailto:christophe.moins@ego-gw.it]
Sent: Monday, July 17, 2006 6:08 AM
To: tech-talk@aps.anl.gov
Subject: CA communication problems

Hello,

I have some problems of communication between an IOC and a client, both
running on different machines on the same subnetwork on a similar
environment (Epics v 13.4.7 under linux redhat 9.0 kernel 2.4.20-8).

At the beginning of the startup of the IOC, process variable values can
be got without problems from a client (either a simple command line or a
php script), but after a while, the data is not available any more and
the following message shows up on the IOC side :

CAS: request from 192.168.101.21:55894 => "bad resource ID"
CAS: Request from 192.168.101.21:55894 => cmmd=12 cid=0xffffffff type=0
count=0 postsize=0
CAS: Request from 192.168.101.21:55894 =>  available=0x1        N=0
paddr=(nil)
CAS: forcing disconnect from 192.168.101.21:55894

Would someone have an idea of what's going on ?

Thanks for your help
Christophe









References:
RE: CA communication problems Jeff Hill

Navigate by Date:
Prev: Re: CA communication problems Kay-Uwe Kasemir
Next: EPICS Vacancies at Diamond Synchrotron Light Source Heron, M (Mark)
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  <20062007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017 
Navigate by Thread:
Prev: Re: CA communication problems Kay-Uwe Kasemir
Next: EPICS Vacancies at Diamond Synchrotron Light Source Heron, M (Mark)
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  <20062007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017 
ANJ, 02 Sep 2010 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·