Vick Khera <vivek at khera.org> writes:
> On Tue, Feb 16, 2010 at 4:30 PM, Christopher Browne
> <cbbrowne at ca.afilias.info> wrote:
>> Perhaps
>> configuration should be pretty much synchronous, checking state against
>> all the nodes, and griping if *any* of them disagree.
>
> If this were the case, you could only ever run configuration changes
> when replication had essentially zero lag.  Ie, all nodes are up to
> date.  If you did, you'd end up waiting until all the events were
> propagated, possibly with locks being held.

Fair enough, that's why I'm not patching anything just yet :-).

There are kinds of changes where I'd expect it to be fine to NOT wait at
all, notably:
 - STORE PATH
 - STORE LISTEN (which isn't used much anymore, of course)

We probably have a bit too much right now that's asynchronous, which
certainly isn't to say that *everything* ought to be made synchronous.
-- 
output = ("cbbrowne" "@" "ca.afilias.info")
Christopher Browne
"Bother,"  said Pooh,  "Eeyore, ready  two photon  torpedoes  and lock
phasers on the Heffalump, Piglet, meet me in transporter room three"