rbowen / centos / board

Forked from centos/board 3 years ago
Clone
Blob Blame History Raw

Attending	1
Agenda	1
Comms out from end of meeting	3
Q&A back to RH stakeholders	4
Attending
Carl
Jim
Johnny
Karsten
KB
Mike
Tru
Regrets:
Ralph
Fabian 


Ref:
1. RH Stakeholder responses to CentOS BOD Feedback
2. New Blog post being proposed 

Agenda
Recap response doc
JH: need to understand the biz constraints that are causing RHT to implement this.
JP: Spoke to Mike McGrath and Stephanie, infra revenue, representing RHEL income has a marked decline, and there is understanding that the biz impact is there.
CT: there may be some way to get some privileged info to Board
KW: Spoke to Chris Wright last week around time to action
Chris is working on a ver of data that can be shared
MM: question if this change will impact/change the numbers ( re: material harm )
MM: is there any counter proposal that will ‘fly’ - it feels like there is very little compromise here.
KW: depends on the msg they want to deliver
Variations on ‘asking for more time’ - allow us to message the change.
KW: if the problem is that the rebuild needs distance from RHT, then maybe Software Freedom Conservancy umbrella or something can be used to create distance.
putting it under a foundation is not ideal but it may be less distasteful than the immediate killing of the clone
KW: we be open and transparent, communicate what RHT asked for and our agreement
KB: can we provide solutions for use cases that people we care about actually do?
KW:  one trick is that out of the gate, people think of their use case as "it's like RHEL" in terms of the instant fear/flight response to news if you can get them to look at their real use case, then we can address those
Shrink main CentOS with Stream as a SIG?
Why shrink core? Reasons reasons reasons
Leave PHP alone but change Java, etc.
JP: could use some breathing room to release what is & start time line to figure out technical change - should, can, etc.
KW: then we don’t try to ship a broken thing
JP: fedora server goes away and the story unifies between fedora -> rhel -> centos
JH: how does redhat get a respite from the ‘clones’
JP: 
Noone trusts Oracle ( ~ seen as a smaller threat )
We are not trying to compete with the clones, because we can take a path and build the onramp through here
CT:
Dynamic is driven by supply and demand; hosting / sci / etc where CentOS built the user base; RHEL user minimal overlap with CentOS; but the brand is being used by the enterprise to threat RHEL. The user base, eg, hosters, will never be in that place.
OKD has a potential working model
Can we satisfy the existing CentOS user base - without a direct association and linkage with RHEL ( focusing the demographics that RHEL does not productise, keeping the footprint away from where RHEL is going to want to renegotiate )
Oracle doing this not going to matter 
There might be segments beyond hosting / web-services that we may miss, so in the announcement open the doors, make it inclusive for them to come engage.
JH: how close to RHEL we can keep CentOS
JP: short term, drop CentOS8 with 8 kernel, 8.1 user space
Longer term: CentOS would be -devel branch 
JP: There is internal push to have RH Layered Projects to use CentOS Streams
KW: Strong msg from internal projects is pretty strong ( and will support the initiative )
Tru: SL, is trusted - unlike Oracle, to do this work in the public
JP: RHT will engage with SL, to make sure they are fine with things
They are asking for time to transition, given that, they would be willing to work with us.
CERN / FERMI : identified people to talk with.
Tru: what about hardware vendors
The KABI / Whitelist would be ahead of CentOS 
Nvidia drivers as an example, are specified against specific RHEL kernel versions
KW: Should we be saying, as a project, that the brand is being mis-used, and use that argument to setup the change coming down
JH: cve’s etc still need to be included in the content / code. It might land earlier (?)
JP: the bugfix work happens with us, and that means the content lands ahead.
CVE work is embargoed - and should land with us in source at roughly the same sort of time as it lands in product, or soon after
For Karsten to take back to RH stakeholders
board understands the problem
wants to make sure the community comes along, doesn't create a bad response for RH and individuals involved, once done can never come back - new coke, classic coke -- 
TH + CT- no matter how well planned, if it goes wrong, do we have any idea how we’re going to respond
board is working through this as a collaboration for solving the problem
few concerns talking through, tweaks/suggestions, clarifications we need to work through, we'll come back






Comms out from end of meeting
Hi Chris, Mike:

As promised earlier with Mike et al, here is a quick update from our continued CentOS BOD discussions today. Carl was able to join us and get his perspective into the conversation.

The BOD understands the problem being presented by Red Hat stakeholders. The Directors want to make sure the community comes along with project changes, and that it doesn't create a bad response for Red Hat and the individuals involved, which -- once done -- can never be undone.

Basically the Directors are thinking of the story of New Coke.[1] We want to introduce “New CentOS” and are concerned that how we treat “Classic CentOS” may end up with a negative backlash that undermines “New CentOS” before people even give it a taste.

The Board is working this as a collaboration for solving the problem. We have a few concerns we're talking through, and will have some tweaks and suggestions to the make, as well as clarifications to work through.

We're meeting again on Wednesday after a few days of async collaboration (time TBD) and I'll contact you following that.

[1] https://en.wikipedia.org/wiki/New_Coke

Best regards, 


Q&A back to RH stakeholders
How do we know the plan is successful?
What happens if CentOS Stream is too successful and customers start preferring it to CentOS?