EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  <20072008  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  2006  <20072008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: RE: same PV on multiple IOC's - side effect
From: "Jeff Hill" <[email protected]>
To: "'Heinrich du Toit'" <[email protected]>, "'TechTalk EPICS'" <[email protected]>
Date: Thu, 27 Sep 2007 09:46:20 -0600
> I have 2 network cards in my PC - this is to connect directly to both
> the campus and the control network here at iThemba LABS.
> 
> Now I have eth0 and eth1
> 
> If I start an IOC and then look for a PV then both interfaces seem to
> return then I get error, although this is the same IOC?
> 
> CA.Client.Exception...............................................
>     Warning: "Identical process variable names on multiple servers"
>     Context: "Channel: "long0", Connecting to: 196.21.126.246:5064,
> Ignored: aglar.local:5064"
>     Source File: ../cac.cpp line 1208
>     Current Time: Thu Sep 27 2007 12:08:45.784685539
> ..................................................................
> 
> aglar.local = 172.16.4.264 - my control net IP.

In this situation, with two network interfaces installed and enabled each
attached to a different subnet, the server will listen on both subnets, and
the client will search on both subnets, by default. So the client library
will find the PV on the same IOC twice in that situation and it is unable to
choose which one is best because the responses come from different IP
addresses. This much ambiguity requires some sort of warning message.

With the portable server we can set EPICS_CAS_INTF_ADDR_LIST and configure
the server so that it will listen on only one of the subnets, but
unfortunately the original rsrv server is still used in the IOC in R3.14 and
it does not yet allow this type of configuration (via
EPICS_CAS_INTF_ADDR_LIST).

However, if it _is_ appropriate for this server to listen on both subnets,
then this is actually only an issue for clients that must run on a
multi-subnet attached host that is also hosting an IOC. For such clients we
do have a solution. You can set EPICS_CA_AUTO_ADDR_LIST=NO and set
EPICS_CA_ADDR_LIST to the broadcast address of the subnet that clients need
to search on. With R3.14 you can choose no more than one subnet for clients
running on a multi-subnet attached host that is hosting also an IOC.

In R3.15 we should add support for multicasting so that it is easier to
configure EPICS systems that live on multiple subnets, and maybe a DNS based
(or some other hierarchical) name resolution option might also be nice for
really large systems. 

Jeff

> -----Original Message-----
> From: [email protected] [mailto:tech-talk-
> [email protected]] On Behalf Of Heinrich du Toit
> Sent: Thursday, September 27, 2007 4:12 AM
> To: TechTalk EPICS
> Subject: same PV on multiple IOC's - side effect
> 
> This is interesting - not a bug really. But maybe somebody knows how to
> stop the error message?
> 
> I have 2 network cards in my PC - this is to connect directly to both
> the campus and the control network here at iThemba LABS.
> 
> Now I have eth0 and eth1
> 
> If I start an IOC and then look for a PV then both interfaces seem to
> return then I get error, although this is the same IOC?
> 
> CA.Client.Exception...............................................
>     Warning: "Identical process variable names on multiple servers"
>     Context: "Channel: "long0", Connecting to: 196.21.126.246:5064,
> Ignored: aglar.local:5064"
>     Source File: ../cac.cpp line 1208
>     Current Time: Thu Sep 27 2007 12:08:45.784685539
> ..................................................................
> 
> aglar.local = 172.16.4.264 - my control net IP.
> 
> mmm
> anyways
> 



References:
same PV on multiple IOC's - side effect Heinrich du Toit

Navigate by Date:
Prev: RE: Question about CA and optimization Jeff Hill
Next: Re: same PV on multiple IOC's - side effect Andrew Johnson
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  <20072008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: same PV on multiple IOC's - side effect Heinrich du Toit
Next: Re: same PV on multiple IOC's - side effect Andrew Johnson
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  <20072008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 10 Nov 2011 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·