the 'eventual consistency' is ok for some use cases, but for those cases you have tools like Cassandra which are orders of magnitude faster. I just don't know who'd want to deal with mongodb consistency model in production.
Agree that eventual consistency is inappropriate for the majority of database workloads. This is why we has always defaulted to strong consistency, by reading from primary nodes in a cluster. (Because eventual consistency is sometimes useful, MongoDB offers eventual consistency, too.)
Agree that eventual consistency is inappropriate for the majority of database workloads. This is why we has always defaulted to strong consistency, by reading from primary nodes in a cluster. (Because eventual consistency is sometimes useful, MongoDB offers eventual consistency, too.)