/home/production/cvs/JSOC/doc/changes_to_dcs_lev0.txt 22Nov2011

Proposed changes to dcs/lev0 processing
---------------------------------------

 Process retransmitted and reprocessed qac/tlm files from a
seperate directory. This will allow the real time stream to continue
undelayed while the retransmitted/reprocessed data is processed.

Change ingest_tlm.c on the dcs nodes:
If the accounting file is a .qacx file, this means it is rexmitted
data and the data will not be moved to the normal backend pipe dir:

/dds/soc2pipe/[hmi,aia]

but to:

/dds/soc2pipe/[hmi,aia]/rexmit

Run 4 additional processes on cl1n001:

ingest_lev0 vc=VC02 indir=/dds/soc2pipe/hmi/rexmit 
ingest_lev0 vc=VC05 indir=/dds/soc2pipe/hmi/rexmit 
ingest_lev0 vc=VC01 indir=/dds/soc2pipe/aia/rexmit 
ingest_lev0 vc=VC04 indir=/dds/soc2pipe/aia/rexmit 

Any data to be reprocessed will be dropped into the rexmit dir.


Proposed changes to dcs .asf file handling
-------------------------------------------
On 2011-11-17 (day 321) we restarted dcs0. At 15:44 (all times UT) we
process the 321_15_00 dsf. We create the 321_15_00 asf also at 15:44.
This is past the 15:30 time that the DDS picks up the .asf file, so the
asf file in not seen by the DDS until 16:30. Meanwhile the 321_16_00 dsf
file comes in and it contains the files that we already acked in the
321_15_00 asf. Now we think these files were acked, so we have no more 
record of them, so we ask for a rexmit of these files. 

There are two different ways we can fix this:
1. The DDS sees that no asf file has been processed since the last dsf
and so will not include anything from this last dsf in a new dsf it
is about to make.

2. The SOC sees that it still has an old asf and so will not include
anything in this asf in a new asf it is about to make.