[Linux-ha-dev] Possible code rationalisation
Michael E Brown
Fri, 15 Dec 2000 17:39:19 -0600 (CST)
Hear! Hear! Extremely good point, and well-stated, at that.
On Fri, 15 Dec 2000, Alan Robertson wrote:
> One of the main "bees" in my bonnet, as it were, is creating a set of APIs
> and cluster tools which are usable by heartbeat, and by other cluster
> managers as well. This level of independence is vital, and is not in any
> sense optional. In some respects, it is the point of the whole exercise.
> It is likely that *something* can be done here, but whatever the solution,
> it needs to be something which all the parties using the packages for their
> various purposes can easily live with.
> Whatever we might choose to do to restructure things needs to understand
> that we are building at least three different vital things here:
> 1) A usable set of software, sometimes called "heartbeat" for creating and
> managing failover clusters
> 2) A set of components which are useful in building cluster software of
> kinds, including failover cluster managers like heartbeat
> 3) A community of people who are creating these things together.
> We've been trying to put out a stable release for a month or two at this
> point. I don't mind changing build methods at this point, as I believe I
> can quickly verify that nothing is broken. However, I'm unwilling to
> undertake any significant code revisions before the stable release comes
> out. Let's put that off until January. Is that acceptable?
> -- Alan Robertson
> Linux-HA-Dev: Linux-HA-Dev@lists.tummy.com
> Home Page: http://linux-ha.org/