[Cluster-devel] cluster/group/daemon cman.c cpg.c gd_internal. ...

[Robert Peterson <rpeterso@redhat.com>]

teigland at sourceware.org wrote:
> 	Moving the cluster infrastructure to userland introduced a new problem
> 	that we didn't need to worry about before.  All cluster state now exists
> 	in userland processes which can go away and then come back like new, i.e.
> 	unaware of the previous state.
Hi Dave,

You know this new development cman stuff and I really don't, but I was 
just thinking:

If we used a shared memory segment, we could hold state information 
there and
then cman would remember the cluster state after process termination and 
restart,
possibly making this whole thing unnecessary.  Just a thought.
Of course, one could also argue that if the process terminated, can we 
really trust the state
information it had at the time?

Bob Peterson
Red Hat Cluster Suite