#opnfv-cperf: cperf

Meeting summary

1. 1. No word from Veryx folks about status of PktBlaster, last info was after dfarrell07 sent docs about how to add tooling/tests (dfarrell07, 18:16:34)
   2. Nikos sent dfarrell07 an update via email, team has been on holiday and he has had family health stuff to deal with so nothing much to report (dfarrell07, 18:17:59)
   3. dfarrell07 has been pulled to high prio downstream packaging stuff, not much cperf this week :( (dfarrell07, 18:18:53)
   4. jamoluhrsen reports that he's making progress on getting Apex running (dfarrell07, 18:19:38)
   5. jamoluhrsen is now able to manually run job on jumphost, so easier to deploy/debug (dfarrell07, 18:19:51)
   6. jamoluhrsen is hitting issues with some Apex docs not being detailed enough for bare metal, is talking to Tim and Dan (dfarrell07, 18:20:13)
   7. Al went to a talk from person working on ODL deployment designs, looking at 3 controller clusters and remote cluster support (2 in 1 site, 1 in another) (dfarrell07, 18:21:13)
   8. Hammers home importance of getting cluster stuff solid (dfarrell07, 18:21:41)
   9. jamoluhrsen will have clusters deploying in apex soon (dfarrell07, 18:21:49)
   10. Al asks about POD shutdowns, jamoluhrsen hasn't seen yet, has been using np (dfarrell07, 18:22:10)
   11. dfarrell07 has ping from Chris W/ODL Board about ideas around getting LF project infra more organized cross-project (dfarrell07, 18:23:57)
   12. LuisGomez talks about clustering, having primary site in local data center that does voting (for speed), and have secondary cluster site that can take over but doesn't normally vote (for speed) (dfarrell07, 18:24:50)
   13. LuisGomez reports that this^^ isn't tested today, currently 1 site with 3 nodes (dfarrell07, 18:25:09)
   14. Al says they are interested in this down the road, so we need to start paying attention (dfarrell07, 18:25:31)
   15. idea with LF infra is to better use resources, maybe by having admins work on a tool (gerrit, Jenkins, etc) for many projects, vs knowing many tools for one project and having many such people (dfarrell07, 18:26:31)
   16. another idea, relevant to cperf, is to share access to hardware (dfarrell07, 18:26:48)
   17. so may be way for getting access to hardware labs in ODL, or more easily doing cross-project testing, especially on complex hardware (dfarrell07, 18:27:18)
   18. LuisGomez talks about ODL should create recommendations for how to deploy clustered solutions (dfarrell07, 18:28:00)
   19. LuisGomez has been out having a baby. Congrats! (dfarrell07, 18:30:27)

Action items

1. (none)

People present (lines said)

1. dfarrell07 (21)
2. collabot` (3)