Date: Wed, 17 Dec 2003 16:24:40 -0600 (CST)
From: Marco Rescigno 
To: Takasumi Maruyama 
Subject: how to sync b0svt06
In-Reply-To: <200312172213.hBHMD2wc1302688@hep.uchicago.edu>
Message-ID: <Pine.LNX.4.44.0312171617000.7324-100000@nroma9.fnal.gov>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

I Taka,

 if you login to b0svt06 you can type:

b0svt06> svtGfr 1 

and the program will only react to GLOBAL_FREEZE coming through G_LINK
directly from b0svt07 every 60 sec or so. In this case 
we will serve very little statistic to the beam fitter but we don't care. 
If Decision Time Out happen you have to disable the auto HRR and leave the 
system in HALT. After at most 60 sec you will have a GLOBAL_FREEZE and
a spydump. Then you can also restart DAQ. 

Looking in $SVTMON_DATA_DIR/spydusmps

you should be able to find 8 files with approximately the same timestamp
in their filename. All of them should have on the first two line 
something like this:

NS 39994 2003/12/17 16:20:45 2 (0) INIT 0 ERR 0 FRZ 1 LLOCK 0 GLOBAL_FREEZE
RC 176570 HALTED 0 28.30 F 119299880 M 114462088 N 791

(i.e. GLOBAL_FREEZE and RC status == HALTED)

in this case you should expect that every buffer in every files should 
finish with the same EE. If this is not true note it down in the e-log, 
it may mean something.... In particular I would check the EE from the 
HF (that you find in the MERGER INPUT B/C/D dumps) and the TF EE and 
the GB EE to look for possible mismatches. We should also 
ask the L2 colleague to dump also the BC Id in their monitor so that 
we make sure wich event L2 is expecting and which one is being processed 
by SVT instead.


   Marco 


On Wed, 17 Dec 2003, Takasumi Maruyama wrote:

> 
>    Hi, Marco.
> 
>   We are having beam, but still only have no SVX data.
>  If possible, can you tell me how to slow down b0svt06
>  spy data acquisition?
> 
>   We are not sure we have SVX until Sunday when you plan to 
>  leave here. (though most likely we have)
>   
>   Regards,
>   Taka.
>