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  <20092010  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  2007  2008  <20092010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: RE: choosing snmp version
From: "Rees, NP (Nick)" <[email protected]>
To: "Euan Troup" <[email protected]>
Cc: [email protected]
Date: Tue, 30 Jun 2009 09:24:48 +0100
Hi Euan,

I have also had a hack at devSnmp (but I didn't get terribly far).

My issue was that I had a problem when it sends out snmpGet requests -
my version seemed to lose a lot of transactions at higher load levels,
which the example program doesn't. I poll 36 devices every 10 seconds
for about 3 values each and the 36th device often doesn't respond. If I
poll every 5 seconds only the first 10 or so devices respond. If I
understand this correctly, under the hood it is actually continually
polling waiting for an answer, and the 10 second timeout just means that
it got no response in 10 seconds, despite multiple tries. However, the
example program can poll all devices reliably in a few 10ths of a
second. I think it is because the UDP messages are getting lost because
of the way the driver synchronizes the get's, and the sample program
avoids this by serializing the messages slightly.

In looking at this, I replaced a number of the helper libraries with the
standard EPICS equivalents during the analysis phase, so the amount of
code is now much less, but I must admit I never fixed the main problem
because polling slower was a work around and I didn't get back to fix it
properly.

However, reading your slides, it seems your changes are much more
ambitious and may also address the problem. However, we don't like to
run as root very much, so is the trap handling an essential part of your
code, or can it be ifdef'ed out whilst still preserving other benefits?
Some of the restructuring seems along the lines I was thinking.

I wouldn't mind having a copy of your code to see how it works, and I
certainly think we should work on getting a new devSnmp release out.

Cheers,

Nick Rees
Principal Software Engineer           Phone: +44 (0)1235-778430
Diamond Light Source                  Fax:   +44 (0)1235-446713 
-- 
Scanned by iCritical.


References:
Re: choosing snmp version Euan Troup

Navigate by Date:
Prev: reading with the asyn record Patrick Thomas
Next: Re: reading with the asyn record Eric Norum
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  <20092010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Re: choosing snmp version Euan Troup
Next: reading with the asyn record Patrick Thomas
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  <20092010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 31 Jan 2014 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·