CloudServices/Meetings/2009-11-10

From MozillaWiki
Jump to: navigation, search

Agenda:

Follow-up items:

  • AI: We should document all the scenarios we are interested in testing and prioritize them for 1.0 (Beta/GA) [MHanson]
  • AI: We should update our capability model docs with more accurate numbers. [MHanson]
  • Split user reg and sync server code into separate repos. (Toby)
  • Need load tests to work against master-only config (mhanson)
  • Need a client that points at the staging 1.0 server (mconnor)
  • mailman list for people running their own weave server (mconnor)


Beta:

  • Client
  • Server
  • Operational readiness

1.0:

  • Client
  • Server
  • Operational readiness
    • No replication architecture for DBs and failure modes bug 526084


Notes:

Follow-up:

  • We did create load tests, but the AI is still open to list all the
  • Capability model has been updated, but it is an ongoing thing.
  • Zandr will need to run the read/write mix tests in master only mode.
  • bug 526903 has been created for Toby's repo separation AI. Zandr will follow up on this.
  • AI: Mardak will spin a custom build with the 1.0 URI for the server tests.
  • AI: We need to do integration tests.
  • mozilla-labs-weave-server-users has been created on google groups.

Beta:

Client:

  • We need UI for Starting Over (backup restore). There should be an RC build out tomorrow and release the client on Thursday.
  • Other issues from the forums that need bugs
    • Clear logs
    • History sync (number of items as well as sites only visited once)
    • Direction of sync (wipe local/remote)
    • Recover secret phrase (inform users that it is stored on their other computer)
    • Identity clients by using host names
    • Detect if noscript is present and ask users to turn it off while doing weave account registration.
  • Issues that should be addressed already
    • Jumbled bookmarks - shouldn't be an issue with


Server:

  • 1.0 API is done.
  • Server is going to be pushed out today.
  • We have not brought up any additional hardware, but zandr thinks we are OK with current hardware based on our load test numbers.

Operational readiness:

  • We need to document our current capacity, expected failure modes and corrective actions to take on the wiki. Especially since Zandr is going to be on the road on Friday and Monday.