Hi, >>I don't see a need why the parent node needs to be updated >> when a child node is added, removed or updated. > >However, getting a consistent journal across >cluster nodes seems more difficult here as you said.
One option might be to persist the journal within or near the node data itself. I guess move and re-order operations could be modeled as delete+add (only on the storage level; not in the journal). The node id could be the complete path (which isn't terribly efficient, but avoids problems). This sounds like a promising approach for MongoDB. Regards, Thomas
