Better handling of NotReady error

TODO

@@ -34,6 +34,10 @@
         - Empty storages rejected during recovery process
       Masters implies in the election process should still reject any connection
       as the primary master is still unknown.
+    - Wait before reconnecting to a node after an error. For example, a backup
+      storage node that is rejected because the upstream cluster is not ready
+      currently reconnects in loop without delay, using 100% CPU and
+      flooding logs.
     - Implement transaction garbage collection API (FEATURE)
       NEO packing implementation does not update transaction metadata when
       deleting object revisions. This inconsistency must be made possible to

neo/lib/handler.py

@@ -174,10 +174,14 @@ class EventHandler(object):
         except (AttributeError, ValueError):
             raise UnexpectedPacketError(message)
 
+    # XXX: For some errors, the connection should have been closed by the remote
+    #      peer. But what happens if it's not the case because of some bug ?
     def protocolError(self, conn, message):
-        # the connection should have been closed by the remote peer
         logging.error('protocol error: %s', message)
 
+    def notReadyError(self, conn, message):
+        logging.error('not ready: %s', message)
+
     def timeoutError(self, conn, message):
         logging.error('timeout error: %s', message)