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: ChannelArchiver update
From: Kay-Uwe Kasemir <kasemirk@ornl.gov>
To: tech talk <tech-talk@aps.anl.gov>
Date: Mon, 27 Mar 2006 13:46:20 -0500
Hello:

There's a new version of the archiver toolset,
accessible from the links on the APS web page.

The web page was actually already updated last
Friday, but that snapshot includes an error.
So if you fetched an update on Friday,
try again today.

There are tree types of changes:
- "Example Setup" chapter and the scripts to support it
  (ArchiveDaemon, ...) have been modified.
  Before, we tried to periodically update the 'master' indices
  e.g. every 20 minutes. That no longer worked with the amount
  of data we have at the SNS (~3..5GB per day).
  Now we separate the 'sampling' computers (running archive engines)
  from the 'serving' computers (indexing, serving).
  This way, we can access 'all' the data for 2005 & 2006,
  but the catch is: There's a split between
  'all' referring to all the currently running archive engines
  and 'all' referring to all the older data
  (typically: Jan 2005...yesterday).
  So it's not perfect, but it's the best we have for now.

  If you liked the ArchiveDaemon.pl as it was:
  Assert that you keep the older copy around,
  don't overwrite it with the new one.

- EPICS base R3.14.8.2 introduced an 'assert' that might stop your
application when the nanoseconds of the time stamps are not
normalized, i.e. exceed 1e9, and you convert between epicsTime
and epicsTimeStamp.
Jeff recently added a patch that makes this more consistent,
so your application is very likely to get an exception when
trying to use non-normalized time stamps.
That's good, except some IOCs at the SNS still occasionally
have such bad time stamps because of ongoing timing system development.
So the archiver code tries to catch these time stamps.


- Archiver code uses C++ exceptions, which sometimes leads to better
  error messages, since for example a low-level read error exception
  gets passed up and printed together with the channel name or
  other context information.

Thanks,
-Kay


Navigate by Date:
Prev: RE: AlarmHandler as a daemon or service Ernest L. Williams Jr.
Next: gateway enum writes Leicester, PJ (Pete)
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: AlarmHandler as a daemon or service Liyu, Andrei
Next: gateway enum writes Leicester, PJ (Pete)
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 ·