====================== #opnfv-infra: infra-wg ====================== Meeting started by bramwelt at 15:03:21 UTC. The full logs are available at http://ircbot.wl.linuxfoundation.org/meetings/opnfv-infra/2019/opnfv-infra.2019-10-21-15.03.log.html . Meeting summary --------------- * Agenda Bashing (bramwelt, 15:03:34) * LINK: https://wiki.opnfv.org/display/INF/2019-10-21+Infra-WG+Meeting (bramwelt, 15:05:18) * One central place for meeting details (bramwelt, 15:06:58) * Confusing when there are multiple places for meeting agendas/schedules (bramwelt, 15:07:22) * LINK: https://wiki.opnfv.org/display/INF/Infra+Working+Group (jentoio, 15:07:35) * LINK: https://wiki.opnfv.org/display/meetings/Infra+Working+Group+Meeting (jentoio, 15:07:48) * Jack plans on running meeting initially (bramwelt, 15:09:21) * AGREED: Meeting details will stay in INFRA space and be linked to the meeting space https://wiki.opnfv.org/display/INF/Infra+Working+Group (bramwelt, 15:12:18) * Jack plans to move the meeting notes under INFRA (bramwelt, 15:13:52) * LINK: https://wiki.opnfv.org/display/INF/Meeting+notes (bramwelt, 15:13:54) * Discussion on how to manage and organize meeting minutes (bramwelt, 15:15:17) * ACTION: bramwelt to clean up minutes to be in individual pages (bramwelt, 15:16:23) * Jack moving minutes from Meeting space to Infra space (bramwelt, 15:16:36) * ACTION: Jack to clean up the infra-wg page under the Meeting space and point to the agenda (bramwelt, 15:19:42) * Discuss agenda (bramwelt, 15:21:13) * Pharos (bramwelt, 15:22:45) * Jack plans to reach out to community lab owners and get a status update (bramwelt, 15:22:57) * Will try to encourage lab owners to come to infra-wg (bramwelt, 15:23:17) * Jack also planning on cleaning up Pharos JIRA (bramwelt, 15:23:31) * Infra wiki needs to be cleaned up / archived / migrated (bramwelt, 15:23:53) * Lincoln mentions Pharos spec being updated, LaaS gives you one-click-openstack (bramwelt, 15:25:06) * What is the outlook for the infra projects? (bramwelt, 15:25:31) * Jack mentioned Parker should be part of infra-wg given the nature of LaaS (bramwelt, 15:25:56) * Also mentions Pharos direction towards PDF/SDF and refresh of Pharos specification (bramwelt, 15:26:47) * Lincoln mentioned discussions at CNTT around Pharos spec and PDF (bramwelt, 15:27:16) * There wasn't take up in the structure and content of the PDF/SDF previously, and the hope is CNTT/RI will help drive and align that content (bramwelt, 15:28:34) * PDF/SDF very helpful for Lab owners as it provides a single description of the environment and makes it helpful to determine what the infrastructure looks like (bramwelt, 15:29:27) * Jack explains context of SDF being collection of components needed for the scenario - CNTT will help set what those scenarios should be and determine the selection of components (bramwelt, 15:30:40) * PDF defines hardware in OPNFV configurations (RAM, Disk, CPU, NICs) (bramwelt, 15:31:22) * SDF defines deployment configuration (networks, IPs) (bramwelt, 15:31:54) * Lincoln suggest RI is it's own scenario - Consumability of that scenario for XCI or Airship shouldn't differ - That part of the SDF was missed before (bramwelt, 15:33:39) * Discussions still needed on implementation and consumability of SDF (bramwelt, 15:34:05) * Jack hasn't been able to attend F2F where the SDF may have been discussed to provide feedback (bramwelt, 15:35:02) * Michael asks if Pharos or Infra should be driving those specifications? (bramwelt, 15:36:47) * Jack states it hasn't been clear where that should happen as the discussion have taken place outside those projects (bramwelt, 15:37:38) * Lincoln suggests Infra helps provide/locate hardware for projects to help meet needs for reference implementation - LaaS doesn't care so much about the format, but the agreement on the content (bramwelt, 15:38:53) * More requirements need to be gathered from CNTT to determine new specifications - networking infra (10/25g?), size of machines, etc (bramwelt, 15:40:44) * Mark mentions '10 server' was not a random selection. (bramwelt, 15:42:07) * Jack explains how process around Pharos specification may work in the future - installer -> hardware -> deploy -> testing -> ovp (bramwelt, 15:44:16) * Pharos spec wasn't able to get to 'sizing' in the past (bramwelt, 15:45:26) * Discussion on LaaS being used to help determine specification (bramwelt, 15:46:02) * Lincoln explains dimensions of flexibility in LaaS comes from the network configurations, but not hardware (currently) - Machines are static hardware (bramwelt, 15:47:25) * Jack explains LaaS is closer to cloud hardware (bramwelt, 15:48:38) * All labs except LaaS at UNH are static (bramwelt, 15:50:41) * Lincoln explains how labs may transition from static to dynamic LaaS (bramwelt, 15:51:34) * LaaS vision is that given a PDF -> hardware is returned configured as specified, and given SDF -> servers are configured to match the scenario needs (bramwelt, 15:52:59) * LaaS goal is to provide community way to provision hardware easily through the UI (bramwelt, 15:54:37) * Mark asks what the overlap between installers and LaaS is? (bramwelt, 15:57:37) * Lincoln answers there is 0 overlap (bramwelt, 15:57:48) * Installers are still needed to provide a deployment of OpenStack (bramwelt, 15:58:40) * LaaS does not provide the VM images need to do the deployment either (bramwelt, 16:00:31) * Jack mentions LaaS, Pharos, and infra projects have been involved in the space for a while and it's great to have more people involved and interested (bramwelt, 16:01:20) * Mark asks why each POD has individual IP? (bramwelt, 16:03:15) * Jack explains it's a benefit to reduce confusion around address space (bramwelt, 16:04:10) Meeting ended at 16:04:56 UTC. Action items, by person ----------------------- * bramwelt * bramwelt to clean up minutes to be in individual pages People present (lines said) --------------------------- * bramwelt (57) * collabot (5) * jentoio (2) Generated by `MeetBot`_ 0.1.4