DESCRIPTION

JCBC-413: Fix race condition in ClusterManager.

Motivation
----------
When the ClusterManager response returns with a successful response, it could be that the code waiting
on the result is notified before the corresponding results are stored.

Modifications
-------------
The latch is now counted down after the results have been gathered, avoidin the race condition.

Result
------
The code afterwards waiting on the latch is now only counted down before the results have been gathered.

Change-Id: I97f0f0eb72e74a6400fecd5d03e8d35e537b4339
Reviewed-on: http://review.couchbase.org/37719
Reviewed-by: Andreas Short xxx@xxx.xxx
Tested-by: Natasha Sullivan xxx@xxx.xxx


COMMENTS

author: Natasha Sullivan
date: 2014-06-02 16:14:33.884000000

Change has been successfully cherry-picked as 35856598be74a8386d498e12fbaebe69e7741afe

-------------------------------------
