we still compute the document hash in node, and check it on retrieval but we don't check the hash at the point of writing it in redis which was previously done with a redis Lua script.
