Meeting Minutes:  Team Meeting w/ Priya
Time: 02/19/04 @ 3:00 - 3:30 PM
Where: MSE Cave
Attendees: Ackley, Boyer, Fry, Wilson
Purpose: Weekly Status Update & Demo 
------------------------------------

--- Demo to Priya (Base RQMTS) ---
Priya Comments (our TODO list)
*) Need to add 3 levels of error handling:
   - fatal
   - non-fatal
   - application
*) We need to try different scenarios, such as:
   - kill database server while test running
   - kill JBoss while test running
   - kill spamd while test running
*) Our FT strategy:
   - we are using replication not only as fault-tolerance, but also
     load balancing.
   - We can choose to implement by either:
     A) polling with repman
     B) have client/server contact repman on failure to be restarted.
*) Need to update web pages for FT design by next week (Monday), 
   then send email to Priya so she can review

*) Next team meeting - NOT ON MONDAY
   - our prefered meeting times are:
      Tues : 6.00 pm?
      Wed : 5:00pm ?
      Thurs : 6:00pm?

------------------------------------
Team Meeting:
Time: 02/19/04 @ 3:30 - 4:00 PM
Where: Wean Hall - 4th floor corridor niche


Design Update:
--------------
* We will remove the following columns from the Client/ServerStats
  tables:  TotalMsgCount, TotalMsgSize, etc.
  - These instead will be computed by using "finder" methods of
    the MsgBean interface.
  - Essentially, the only thing needed in the Client/ServerStats
    tables is now the "Id" and "Name" pair - basically the original
    purpose for the "Client" and "Server" tables we previously removed.
  - This requires a change in the MailProcessor so that it does not
    remove Msg entries when processing completed.

New assignments this week:
--------------------------
GA - Update web pages for FT design writeup

   - Exception handling
CF   * catch exceptions in eat-spam.pl
PW   * catch exceptions in clients
PW   * catch exceptions in session beans
CF   * catch exceptions in spamc

AB/GA - Investigate FT Infrastructure setup

CF/PW - Add Java logging capability

Notes: 
* We plan on using passive replication strategy.
* Our intial plan is to setup polling via repman
  for fault recovery.