Date and Time: 6:00am PDT / UTC 13:00, Thursday August 20, 2015
Convener: Bin Hu (AT&T)
Adi (Huawei) introduced Pinpoint proposal. It is a requirement project for fault localization, i.e. a process of deciding the exact source of a failure from observed indicators.
Chris P (Ericsson) asked if those diagrams already exist today, or will the project plan to build it. Adi answered that it is a requirement project, including gap analysis and BP if gaps are found. There may be a PoC but not a complete implementation.
Dave N (RedHat) asked what is the relationship between fault correlation and fault localization. Adi answered that Doctor project talked about "fault correlation" but it actually means the improvement on BP. This proposal is doing "fault correlation" by finding the root cause of fault from related faults.
Ildiko (Ericsson) indicated that Doctor is moving step by step. Although it is not there yet but its scope might cover fault correlation. She suggested to talk to Doctor project. Adi said that he would be happy to talk to Doctor project, and collaborate with them.
Adi will take the action to:
- initiate discussion with Doctor project in mailing list
- contact Doctor project lead and present it at next Doctor project meeting to get alignment
Adi will report the result at next meeting.
Daniel (RedHat) introduced this project proposal.
Chris (Ericsson) asked how could it run in OPNFV. It needs further discussion with test projects.
Bin suggested that we recommend it for TSC Creation Review, while we suggest them to have further talk with test team for details of integration with OPNFV test framework.
Rob (from ARM) introduced this proposal.
Ana (Ericsson) asked if this project will contribute specific test cases to test projects. Rob said that it is something they will explore when doing this project. And they intend to reuse current generic test cases, and will contribute ARM-specific ones if any.
Chris asked if we need a project or we can do it in a different way, e.g. community group. Rob indicated that in an ideal way, every component will work nicely.
Bin indicated that from a service provider and end user perspective, this project is important to provide an alternative option for OPNFV to run a ARM servers. When we consider the way of doing this work, we also need to consider to prevent from extra work and negative impact on other projects. We want to have this to be part of B release, while minimize any risk.
The proposal will be revised to address above concerns.
Time runs out. Postponed to next meeting.
Inactive and postponed until author wishes to resume discussion