Preview Countdown Meeting - April 23, 2007

Agenda

• Review of schedule and upcoming deliverables
• Cosmo 0.6.1 migration (Ted)
• Test data (Katie)

Notes

Schedule and Milestones Update

• Desktop
  • Performance work progressing
  • Revise PPC numbers but there is still some work to do
  • We are considering the possibility of extending the performance work period
  • We may have to fix some bugs that block testing - PPD and QA will work on a list this week
• Cosmo - revising the 0.6.1 dates
  • Ted * Not tracking the interop bugs on the Chandler side - should do this * Tracking the Cosmo bugs * From what we can tell now, we are looking at longer cycle to get all of these fixed
  • We are starting to send out regular Cosmo status messages updating everyone on the progress
  • QA sessions: Easier to just log the bugs and have Morgen look at the bugs. Morgen can't always respond in real-time.
  • There are some issues around what machine we are using to test against
    • Thinking of moving to Randy's machine rather than QACosmo
    • Bear concerned about load when we have many testers
    • Turn around for checkpoints is too slow - Randy needs to hook up a debugger
    • Ted, Aparna, Randy, Bear will discuss appropriate machine to use for this testing
  • QA next steps
    • Email users list - ask people to send in calendars
    • Email dev lists and have people try and publish
• Production hardware
  • Production hardware for died
  • Jared spent some time trying to reinstall, didn't work.
  • Box is under warrantee and Jared will look at the next steps for dealing with that

Cosmo 0.6.1 Data Migration

• Mikeal proposed not upgrading osaf.us and putting Cosmo 0.6.1 on a new server, have people move over manually when they are ready
• Jared - some concerns
  • Some of the arguments don't seem valid ie: supporting Alpha4. We probably don't really need to support Alpha4.
  • We could migrate some data ie: account data but not critical
  • We have about 100 users currently
  • Zero day breakages with this proposal - leave the old server up
  • Would like to see this procedure tested - Philippe tried this already
  • What breaks if we were going to try the migration
  • Likes migration because it gives ervice continuity and support for those 100 people
  • People who are not using Chandler will get the updated version without doing anything new
  • Some people using iCal, Evolution
  • Will people end up with -1 collections - this will be confusing. This would be a reason for not doing the migration
• Going to a new server might be easier for people to understand
• Mikeal was motivated to send this proposal by the past experience with the whole migration test process and upgrade
  • Wanted to simplify things from a QA perspective
  • There might be some ETag issue on Cosmo - server data overwriting the client
• Next actions:
  • Need to call out the specific problems. The thread doesn't detail the technical reasons why this is difficult.
  • Jared - concerned that we are not testing the migration. Not testing the process and qualifying the code that does the migration.
  • Resolve this via the thread - can we make this public.

Test Data

• We are blocked on a fix by Brian Kirsch to replace real email accounts with test accounts
• No response on the thread with the status of this
• Apparently the fix was committed so we can go ahead and test it out then get people using it
• Next actions: Respond to the list so everyone knows the fix was committed

Terms of Service

• Still stalled with legal stuff
• Jared, Katie and Mitch tracking the issue

-- SheilaMooney - 22 Apr 2007