> On 16 Oct 2016, at 3:07 AM, Alan Robertson <al...@unix.sh> wrote:
> Hi Ken,
> I'm perfectly happy to transfer the OCF.org domain to Red Hat - or to
> you personally.
> I would prefer for the repository for the standards to not be tied to
> the existing ClusterLabs source repository. I'm perfectly happy for the
> same people to manage it - but I think it's confusing to say "it's part
> of Pacemaker”.

I don’t think anyone is suggesting that.  Pacemaker << Cluster Labs

Cluster Labs is the umbrella under which many in the wider HA community have 
decided to put their projects (much like OpenStack is also a collection or 
independent projects with a common purpose and message).

The intention is that Pacemaker is just another project and when someone writes 
a better cluster manager it can take Pacemaker’s place in the stack without 
loosing the umbrella.

> In practice, that might be essentially true, but I think
> it dilutes the idea of a standard.
> Life got complicated, but the intent of the OCF was to be a set of
> standards defining a framework (not so much an organization). Since Red
> Had originally declined to participate in the definition effort (but
> were asked to), it made sense for it to be separate. I was pleased that
> they eventually implemented part of the standard (pre-Pacemaker).
> I think a certain minimal level of separation still makes sense.
> Otherwise it's just "pacemaker-compatible". That's not a horrible thing,
> but it's less than a semi-independent framework specification.
> Let me know how I can assist you with whatever you decide.
> You could have just gone your own way, but you chose to include me - and
> I thank you for that courtesy.
>    -- Alan
> On 10/14/2016 03:21 PM, Ken Gaillot wrote:
>> Hello everybody,
>> There has been a lot of talk over the years (including on this list [1]
>> and the ClusterLabs mailing lists [2]) of updating the OCF resource
>> agent standard.
>> The standard is currently used by at least the Pacemaker and rgmanager
>> cluster managers, and the Assimilation monitoring system.
>> OCF as an entity faded out long ago, so there is no formal process to
>> update the standard. OCF started as a working group of the Free
>> Standards Group in 2003, but was already inactive by the time the FSG
>> was absorbed into the Linux Foundation in 2007.
>> Since this list has had very little traffic in recent years, I would
>> like to propose these changes:
>> * OCF could now be considered the name of the collection of standards,
>> rather than an organization.
>> * ClusterLabs [3] (the hub of the Pacemaker community) could take over
>> the role of publishing the OCF standards, with updates taking place
>> through pull requests against the ClusterLabs GitHub repository [4].
>> * Anyone still interested in OCF could subscribe to the
>> us...@clusterlabs.org and/or develop...@clusterlabs.org lists [2], and
>> this list could be closed to new posts and members.
>> I'd like to get feedback from anyone here (especially Alan R. and the
>> Assimilation community, and anyone else who uses OCF outside Pacemaker)
>> on whether that sounds reasonable, or whether anyone has a better idea.
>> Much of this has already happened de-facto, but I'd like to make sure
>> there is a community consensus before proceeding with updating the
>> standard, and hopefully consolidating the various OCF websites/lists.
>> [1] http://lists.community.tummy.com/pipermail/ocf/2014-October/001413.html
>> [2] http://clusterlabs.org/mailman/listinfo/
>> [3] http://www.clusterlabs.org/
>> [4] https://github.com/ClusterLabs/OCF-spec
> _______________________________________________
> OCF mailing list
> OCF@lists.community.tummy.com
> http://lists.community.tummy.com/mailman/listinfo/ocf

OCF mailing list

Reply via email to