Mar 02, 2006
JIM porting to SAM v7
Revised Plan
   
This new plan associates people to tasks. It also adds a couple of tasks that 
were not properly considered before (porting the merging code), which may delay 
the project by 20 days from the initial 2 months.

NOTES:

- Initially sam_client will have the sam products directly. Eventually we'll move
  to sampy. This introduces an extra task that takes about 4 working days, but 
  should save time in the development. The time spent in release cycles to develop 
  the suite is too much: let's get a stable suite of sam products, then will move 
  to sampy.

- How do we test xmldb_client and xml_meta_configurator with python 2.4 ?
-- xmldb_client: test all user interfaces from jim_config
-- check if creating the xml job context uses DOM (from the language instead of 
   froo pyxml as of today)
-- xml_meta_configurator: run jim product tailoring

- Discuss how to move from maintaining two brances of the sam station (poll and 
  callback) to deploying/maintaining a single version.
  The sam client will have both functionalities (poll and callback) that will be
  maintained in the same branch (rather than in two branches as of today).

- Porting the merging code should be considered as a separate task, because it 
  requires more effort than initially considered (13 FTE days). This could add 20
  working days to the initial plan of 2 months.

   
- Port jim software suite to python 2.4.
* Parag will manage this:
Proposal: 
** parag works on xmldb_client
** Sudhamsh helps with xml_meta_configurator 
-- Perform show-stoppers tests (critical porting). 
   Time: (3.0 +- 0.7) (min 1, max 5, most-likely 3)
--- xmldb_client (uses XML RCP libraries)
--- xml_meta_configurator (uses pyxml intesively)
-- Perform tests of the complete suite.
   Time: (3.2 +- 0.5) (min 2, max 5, most-likely 3)

- Implement required commands to sam v7
* Andrew is responsible
Proposal:
** Rohit will wotk on sam get dataset
** Amdrew will work on the station code
-- Port "sam get dataset" command from sam v5.
   Time: (2.0 +- 0.3) (min 1, max 3, most-likely 2)
--- Lightweight consumer in poll mode
--- Lightweight consumer in call-back mode (IS THIS NEEDED?)
-- Add "poll"-mode to the station v6 routing interface (for LCG).
   Time: (4.8 +- 0.8) (min 2, max 7, most-likely 5)

- Port execution site software suite to sam v7.
* Gabriele is responsible for this
-- run_grid_job.py and plug in (jim_job_managers).
   Time: (2.0 +- 0.3) (min 1, max 3, most-likely 2)
-- wrappers (jim_job_managers): check the use of the sam command line commands.
   Time: (1.2 +- 0.2) (min 1, max 2, most-likely 1)
* Andrew is responsible for this
** for testing: can mc_runjob be used for merging (does it work with python 2.4) ?
-- jim_merge reco_merge  (NEW TASK)
   Time: (7.7 +- 1.7) (min 4, max 14, most-likely 7)
-- sam merge (mc_merge)  (NEW TASK)
   Time: (5.5 +- 1.2) (min 3, max 10, most-likely 5)

* Parag will be responsible
** Sudhamsh will do the work
- Port cliet site software suite to sam v7.
  Time: (3.0 +- 0.3) (min 2, max 4, most-likely 3)
-- Port "check job consistency" modules (interaction with db server)

- Use sampy instead of individual products (NEW TASK)
* Responsibility to be defined
  Time: (2.8 +- 0.5) (min 1, max=4 , most-likely 3)

-Integrate system
* Responsibility to be defined: Gabriele et al.
-- Integrate JIM (new sam_client software distribution).
   Time: (1.2 +- 0.2) (min 1, max 2, most-likely 1)
-- [ Contingency:  What if we cannot port JIM to python v2.4?
     Distribute different python versions for JIM and Runjob to the worker node.
     Time: (min 4, max 10, most-likely 6)
   ]
-- Integrate Runjob.
   Time: (14.5 +- 2.2) (min 7, max 20, most-likely 15)
   (ASSUMES:
       o Runjob integrated with SAM for data storing and MC request system
       o Runjob developers presence
   )
-- Integration tests.
   Time: (8.5 +- 0.8) (min 5, max 10, most-likely 9)