Log entry time 21:44:44 on October 04, 2010

Entry number 333649

This entry is a followup to: 333632

Followups:

• 333650 10/04/10 22:06 R. Michaels EPICS truncation due to ET insertions

The problem mentioned in halog 333632 appears to be due to a lack of
the variables actually in the datastream.  (Why ?  maybe the variable
doesn't exist or the server is wrong, and running "caget" by hand 
would show that.  The epics logger can also run in a debug mode with
a print statement ...)

Here's how to tell if there's a problem with the Podd software (THaEpics)
or not:

1) Do a reasonably short run, about 10 min.  To ensure a good efficiency,
make the data rate low (10 Hz) to have a more "purified" sample of 
EPICS data since that's all you care about.

2) Use grep to see if the data exists, e.g. 

grep -ai haVMI3128_3_AI28 /adaql2/data3/dvcs10_6976.dat.0 | more

If the data does not exist, then there is no hope of Podd finding it !

3) If the data exists, run Podd for the ENTIRE run (feasible since
it was short).  Now the "E" tree must contain the data if there are 
no typos in the output definition file.  (Perhaps cut/paste the 
variable from the output definition file to be the argument of grep.)  

4) Remember there are a few substitution rules, like colon (:) is
replaced by underbar (_) in the tree since Root doesn't like colons 
(or arithmetic symbols).  But you can discover what the replacememnt 
is by doing something like this with wildcard * : E->Print("haVMI3128*") 

5) If the data are in the data file but not in the E tree, you have
a legit complaint.