Make sure you didn't miss anything with this list of the Best of the Week in the NoSQL Zone (Mar. 21 to Mar. 27). Here they are, in order of popularity:
We all know that MongoDB is web scale, but how, exactly, does one run MongoDB at scale? This recent article provides a list of ten tips for doing so, each including a brief explanation and outlining potential pitfalls and optimization tricks for each.
Quite frequently our users want to create a unique index on a data set but encounter some form of the following error because of duplicate key value(s). While MongoDB supports an option to drop duplicates, dropDups, during index builds, this option forces the creation of a unique index by way of deleting data.
What we all need is a generic way to run functions over data stored in Cassandra. Sure, you could go grab Hadoop, and be locked into articulating analytics/transformations as MapReduce constructs. But that just makes people sad. Instead, I'd recommend Spark. It makes people happy.
Recently the author had a request for Cassandra-style counters at the mailing list. And as long as he is doing a feature design series, he thought that he could talk about how he would go about implementing this. Again, consider this fiction, he has no plans of implementing this at this time.
At an extremely high level, replication in MongoDB and MySQL are similar. Underneath the covers, though, their implementations are quite different. And in peeking underneath the covers while developing TokuMX, the author learned more about his favorite thing in MongoDB replication: crash safety.