April 2015 Summaries

MongoDB 3.0 introduces compression with the WiredTiger storage engine, which supports multiple compression libraries, including Snappy and zlib, offering varying levels of compression rates and resource utilization. Compression is effective for reducing storage needs by decreasing data size, improving disk I/O throughput, and fitting more data in RAM. The benefits of compression include improved performance and reduced storage costs. To test compression with your data in MongoDB 3.0, you can create a new secondary replica set member or restore your data into a standalone instance to compare the compressed sizes of different options. Compression rates vary across databases, such as Enron, Flights, MongoDB Config Database, TPC-H, and Twitter, depending on data characteristics and cardinality. Capped collections are implemented differently in MMAP storage engines compared to WiredTiger, which may affect comparisons between storage engines.

PyMongo 3.0 is a significant release by PyMongo, a partial rewrite of the Python driver for MongoDB. It aims to improve conformance with other drivers, making APIs more standardized and easier to learn. The new MongoClient class replaces multiple client classes, providing a single interface for connecting to a server, replica set, or set of mongos servers. This change simplifies deployment and configuration, allowing users to easily switch between environments. PyMongo 3.0 also improves responsiveness, robustness, and modernity, with features such as automatic reconnection, reduced latency, and better thread safety. The release is the biggest in PyMongo's history and marks a significant improvement over its predecessor, PyMongo 2.0.

The next-generation MongoDB drivers conform to a standardized server discovery and monitoring spec, which provides a consistent way for clients to discover and monitor MongoDB servers. The spec describes how a driver monitors a topology, including the state of available servers, server types (mongos, primary, secondary), and how it keeps this information up-to-date. It explains the process of discovering an entire replica set given an initial host list and responding to stepdowns, elections, reconfigurations, network errors, or the loss of a server. The spec also describes the behavior of multi-threaded and single-threaded drivers, as well as hybrid modes in the C Driver. It outlines how drivers handle crises, such as the primary server becoming unavailable, and how they resolve these situations. The spec provides unit tests to ensure compliance across different drivers, allowing developers to write applications that can work seamlessly with MongoDB servers.