Meeting notes for 28 Sept. 2005

Present: Joanne Bogart, Ric Claus, Sergio Maldonado, Lester Miller, Jim Panetta, JJ Russell, Tony Waite

Facts, decisions, issues and guesses, in no particular order

Calibration runs will return, as part of data stream, logical_id of LATC master and of LCI binary. Physics runs will return logical_id of LATC master and logical_id(s) or equivalent information for LPA. (In the case of physics runs, this is not sufficient to determine the configuration of the instrument while processing a particular event, but, combined with other bits, it will be.)

Start run command will include file_id (or equivalent identifier for RAM) of LATC master and other files. Also includes ground_id. For calibration runs "other files" may just be LCI. Less clear for science runs.

A single calibration from user or analysis perspective consists of several runs, with different LCI scripts and sometimes different LATC configurations. (Might want to keep same ground_id for different runs within a single calibration.)

Would like to avoid or at least minimize processing of duplicate files. Will have very similar LATC blobs which, when split, are likely to have some identical components. They should be recognized as such and only stored and processed once.

Uploads can occur concurrently with other activity, such as calibration or physics running; need not be while the instrument is quiescent. It cannot occur while the instrument is in Boot or Terminal modes. Given the narrow bandwidth of 1553, we're (LICOS) encouraged to overlap uploading with calibration.

Both MOOD and FMX use MySQL. Questions from Tony:

• Should their tables be in the same (in MySQL sense) database?
• There is and always will be a single instance of the FMX database. What about MOOD?

Online (LICOS directly or via MOOT – TBD) needs to invoke FMX commands which modify its MySQL database by adding and updating rows. There is a potential for interference among users of the unique FMX database.

Managing RAM. This has to be done by the operator; FMX can't possibly keep track since it doesn't know when there has been a reboot. It is possible to ask the instrument what is currently in RAM. Don't want to automate start run to the extent that required files are uploaded since often most or all of them will already be there.

To do

• Come up with a guess at procedure for doing a calibration. Assume to start that files will be uploaded to RAM, but would be well to think about sib case also. Expect this will take several iterations and several people.
• Rework MOOD schema, at a minimum eliminating History table and coming up with some way to bind together multiple configurations which go to make up a single calibration algorithm. Will also make space for logical_id in FSW_input table. (Joanne)
• Work out procedure whereby Lester can register sources (e.g. the xml file which is input to LCI and the various LATC inputs) in MOOD. (Joanne)
• Sergio and Joanne, maybe also Tony, will meet to discuss FMX issues.
• Jim has LATC file splitter mostly in hand; just needs to deal more completely with broadcasts.
• There is some work to be done to regularize access to LATC and LCI files; currently only nfs access is supported. (James)
• Get some example LCI runs through the test bed to verify correct LCI input file construction and mechanics. (Lester and James)

Initial draft: 28 September 2005 Last revised:

J. Bogart R. Claus