when talk distribuited databases, them, must have same information? example imagine table customers
select * customers this query must return same result in database? in other word users must exist in databases? example "user 1" must exist in databases?
and imagine table master-detail example, sale , sale_detail if using sistem , "insert new sales (sale , details)" must insert new sale in databases?
and how transaction works here? or sale-sale_details not must in databases?
how distributed transaction work?
a distributed database doesn't mean data replicated server machines. data replication kinda topology thing.
what distributed database means multiple machines work store , serve data. how story. wiki
a distributed database database in storage devices not attached common processing unit such cpu , controlled distributed database management system
what you're asking how data distributed in distributed database.
usually common distribution method hashbased distribution. calculates hash value against every key or id of value , stores in 1 of nodes. yes data not stored in of servers (hence distribution)
hash distribution ensures data more or less uniformly distributed among database server machines or cluster.
your other question how query served can answered first understanding different nodes of server work find result of query. every database server performs query on own local data set (remember data distributed not replicated) , reply client. client api should intelligent enough accumulate result or connect servers returned reader perform accurately.
