Problem:
--------

In order to synchronize excluded metadata objects, users had to
correct the underlying issue, if any, and then trigger the
synchronization of the objects again. This could be achieved
through discovery of individual tables or a complete
synchronization by virtue of reconnecting the MySQL Server to
the Cluster. Table discovery doesn't scale well with an increase
in the number of tables and MySQL Servers while the latter
approach involves the overhead of a reconnection to the Cluster.

Fix:
----

The excluded objects list is cleared when ndb_metadata_sync is
set by the user. This makes all such objects eligible for
synchronization in the ensuing detection run thus providing a
simple way for users to retry the sync of all excluded objects.

An additional improvement is to remove the validation logic for
retry objects which used to take place at the beginning of every
detection run. Since these objects are of interest only for the
length of time ndb_metadata_sync remains enabled, the list of
retry objects is cleared when ndb_metadata_sync is flipped to OFF
which denotes that all changes have been synchronized. A further
effect of this change is that the retry objects data structure no
longer needs to be protected by a mutex. The data structure
continues to be accessed by the binlog and metadata change monitor
threads but this occurs only when ndb_metadata_sync is enabled and
in serial fashion.

Change-Id: Ie49367c05f6434fa36dabe280d09ad7a4348820c