Node instance counter is not properly set after marshaling

node instance counter is not properly set after read of process instance from data base - it's simply to late set - after reading node instances thus causing:

• new node instance id assigned to node instances after being read from db
• since new ids are assigned to active node instances it might lead to duplicates as the counter is reset - not set properly - before node instances are read

there is another issue as well, that new id from counter is always set, regardless of what was set by the marshaller - causing change of node instance id between reads from db for given process instance.