Notice: This Wiki is now read only and edits are no longer possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.
Lessons Learned from Release 0.3
Aperi Wiki Home Aperi Process Lessons Learned
- Lessons Learned for release 0.3:
- Freeform discussion to understand what worked well and what areas we can improve for the future. Topics to include, but are not limited to:
- requirements and planning
- implementation / code
- information development
- third party components
- integration / build
- test
- availability
- tools; cvs, bugzilla, etc.
- Items already suggested:
- IDVT Build process: During IDVT, we need to have one person responsible for the builds and a quick ‘acceptance’ test needs to be done for each platform.
- Review the IDVT reporting charts after R0.3 to make sure they are conveying the needed information.
- Make use of Bugzilla version field
- New items from this meeting:
- We need to look at the process we use to look at requirements for releases and how we analyze them. (Dave)
- We also need a similar process once the content of release has been determined. How to define the content of each milestone, how we determine testing, etc. (Dave)
- Need to give ourselves more time for testing at the end of the process (Tom)
- We need an established procedure (checkbox) for the final build and test. (Dave)
- From the IDVT standpoint we were very late in the game for code integration and final build. We need to plan a two to three week buffer to absorb problems. This issue has been consistent for the last couple of releases. (Hans)
- Since we are line item driven, what does it mean to be complete with a line item? Need to define this. (Todd)
- We need someone dedicated to reviewing or answering questions for release notes or install instructions. (Chris)
- We need to identify any third party code we are using as part of the DCUT exit so that CQ’s can be investigated and submitted early. (Tom)
- Automated build and test process: We need to put together a self contained test system. Every night the build runs, it gets copied to a location, installed on all platfroms, and basic operational tests are run to indicate build is worthy.
- (related to previous one) We should test builds with the new SAN simulator to know the build is decent. (Dave)
- Risk reduction of new technology – what can we do to help prevent finding issues late in the cycle? We always have a risk with early adoption of a technology. A factor to consider at planning time is the complexity of what we are doing. We need to plan appropriate timing based on complexity into the schedule / risk plan. Identify ‘at-risk’ items up front and have a mitigation plan for them if we run into sever problems. (Tom/Dave)
- We need to distinguish between the core functionality of our offering and the new things in our planning. (Dave)
- SNIA 2006 end user council report - Utilize this to answer some of our questions. Put a link in Aperi Dev (if this is a SNIA item for general consumption) or have someone review and talk to it. (Javier)
- Javier volunteered to take this on.
- Freeform discussion to understand what worked well and what areas we can improve for the future. Topics to include, but are not limited to:
- Lessons Learned for release 0.3: