User Tools

Site Tools


Minutes of Technical Community Discussion on April 09, 2015

  • Date and Time: UTC 13:00, Thursday April 09, 2015
  • Convenor: Bin Hu (AT&T)
  • Participants:
    • Adrian Hoban (Intel)
    • Al Morton (AT&T)
    • Artur Tyloch (Canonical)
    • Brady Johnson (Ericsson)
    • Bryan Sullivan (AT&T)
    • Carlos Goncalves (NEC)
    • Dave Neary (Red Hat)
    • Dick Chen (ZTE)
    • Georg Kunz (Ericsson)
    • Heather Kirksey (OPNFV)
    • Howard Huang (Huawei)
    • Ildiko Vancsa (Ericsson)
    • Jerry Zhao (Huawei)
    • Jie Hu (ZTE)
    • Joe Huang (Huawei)
    • Larry Lamers (VMWare)
    • Maria Toeroe (Ericsson)
    • Mario Torrecillas Rodriguez (ARM)
    • Mike Young (Huawei)
    • Narinder Gupta (Canonical)
    • Pat Vaughan (Intel)
    • Prakash Ramchandran (Huawei)
    • Praveen Kumar
    • Qiao Fu (China Mobile)
    • Rajeev Seth (Sonus Networks)
    • Ricardo Noriega de Soto (Ericsson)
    • Ruan He (Orange)
    • Ryota Mibu (NEC)
    • Ulrich Kleber (Huawei)
    • William Masson (AT&T)
    • Yuri Yuan (ZTE)

Ryota introduced the BP in very details, which was pushed to Gerrit.

Dave Neary had a question about how we relate this BP requirement to the implementation of Nova hardware propagation. Ryota planned to use a generic approach.

Adrian asked what the time bound of "immediate" means. Ryota said within 1 second. Adrian further suggested that we should specify timing, because it impacts implementation and frequency of alarms.

Joe asked a question whether or not we have considered the situation that all compute nodes reboot at the same time, which triggers lots of alarms. Ryota described that we should avoid to add new sequence in existing sequence of Ceilometer. Joe further indicated that if there is no flow control mechanism, congestion will happen.

Ryota agreed to further revise the BP:
- specify the timing of "immediate"
- address the flow control mechanism and congestion in case that all compute nodes reboot at the same time.

Uli and Jerry introduced this proposal and clarified that it will add another installer approach in BGS.

Group wished them the best of luck in working with other installers in BGS in order to find the maximum commonality.

There was technical issue with regard to Brady's phone connection. So this presentation was postponed to next meeting.

Bin indicated the discussion status of Multisite Virtualized Infrastructure proposal in the mailing list. It seems that all concerns were addressed. Dave Neary indicated that he had an impression that there are still concerns outstanding.

Consensus is for Bin to send another polling to reach the consensus in the community about which way to go - either to recommend for creation review in TSC if all concerns were addressed on mailing list, or to continue the discussion on the mailing list to address outstanding concerns.

Meeting adjourned

wiki/tc_minutes_20150409.txt · Last modified: 2015/05/06 20:07 by Daniel Farrell