Bug 6622 - ClusterSingletonService registration race condition
This is a hot fix for a race condition issue which can happen during a service lifecycle.
An application which manages services in not aware when the unregistration is finished, hence
service close is followed by immendiate service registration which may fail consequently.
This condition is handled by reatempting service regsiration after 10 ms.
Also fixes RIBImpl - create TX chain in #instantiateServiceInstance and asynchronously clean operatoinal DS.
Change-Id: I110badc8bfb482a5b7cb6545b882cda2d0c661ae
Signed-off-by: Milos Fabian <milfabia@cisco.com>