eod status 27-jun-2012

[Gustavo Niemeyer]

On Thu, Jun 28, 2012 at 1:21 AM, William Reade
<william.reade at canonical.com> wrote:
> In several cases, we don't handle the situation in which a node change
> and revert happen close together and are processed as a single event by
> ZK; affected watchers are ConfigNodeWatcher, NeedsUpgradeWatcher,
> ResolvedWatcher and PortsWatcher. The impact and likelihood of each of

We've discussed ConfigNodeWatcher before, and my opinion is still the
same about it: the use case of attempting to catch configuration
changes that don't actually change the configuration is too expensive
for how unlikely this case is, and if the case ever does happen the
consumer logic is supposed to cope with it, otherwise it's broken.

I wouldn't mind checking the other cases, though, if you feel that
would be good.


gustavo @ http://niemeyer.net