Core Committer Weekly Interlock - July 10, 2017

Attendees

Agenda

2.13.0 release - release is green, release notes posted. AI Former user (Deleted) should only include bugs that have a fixed version for that particular release. (No NA or previous fixed versions in the release). AI Peter Pan remove or move to bottom of page 1.1. Deprecation banner from release page.
RackHD Redfish API "plumbing" email
1. WSMAN based catalogs, redfish APIs and SMI services
  1. Do we care that WSMAN goes right to the source (not catalog) for log data where IPMI uses pollers/cataloged data? And do the implementations need to align?
2. Email snippet to seed the discussion:
  1. The Mustang team is moving Redfish forward with WSMAN based catalogs (generated with SMI services) via https://rackhd.atlassian.net/browse/RI-131. The RackHD Redfish API is sort of a proxy in to the service.
    The Veyron / Corsair team is moving Redfish API implementation forward with IPMI based catalogs via https://rackhd.atlassian.net/browse/RI-130
    At this point in time, the efforts are starting to come together and even overlap. Let me frame up an example:
    LogServices was developed as part of the initial RackHD Redfish implementation going back ~1-2 years, uses a poller mechanism, and returns log data from the ipmi based catalog. This works for both Dell and Non-Dell servers. As part of RI-130, it was considered complete. The Mustang team introduced LogServices support that conditionally checks if it’s a Dell server and returns WSMAN catalog based log information. Although the Redfish API may remain the same, the plumbing has changed for Dell servers. The fact that there is a Dell server check and it collects data directly from the source as opposed to polling the catalog was a point of contention with Jon. AI Maithri_Ajjampura and Jeanne Ohren to investigate with Brad and arch team on go forward plan. For Dell plan is to do a push of log files, vs poll however can we do the same for non dell vendors (Quanta etc).
Replace Vagrant with Docker - Success is considerably higher,\ (98%) time to run has reduced as well (~13 minutes faster). Where are we at with this? Next steps bare metal regression OS Install testing. AI George Paulos where are we at with the story XXXX ? Should we go to InfraSim Docker deployment? Plan is for InfraSim to run both docker and OVA in parallel to verify stability. AI Former user (Deleted) to implement this trial run of both.