August 2012 Summaries

The DataStax OpsCenter Alerts feature allows users to be notified when specific metrics exceed configurable thresholds. This post demonstrates how to retrieve active alerts using Python via the OpsCenter REST API, assuming alert rules have been set up in OpsCenter and are working with a cluster called MyCluster. The code includes functions for getting URLs from the API, retrieving fired alerts and alert rules, connecting the dots between them, and formatting human-readable messages based on the type of alert rule. This can be used to integrate OpsCenter alerts into existing systems via their APIs.

Cassandra is designed for fault tolerance and availability in distributed systems. When a client makes a request, it may talk to any node in a Cassandra cluster, which acts as the coordinator responsible for routing the request to appropriate replicas. If the coordinator fails mid-request or if a replica fails before the request arrives, the client is in the dark and has no choice but to retry. In case of a replica failure after the coordinator has forwarded the client's request, Cassandra replies with a TimedOutException and provides an acknowledged_by count of how many replicas succeeded. The coordinator can force the results towards either the pre-update or post-update state using hinted handoff, which stores the update locally and re-sends it to the failed replica when it recovers.

Dynamic snitching is a feature in Apache Cassandra that determines which datacenters and racks are both written to and read from. It works by monitoring the performance of reads from various replicas and choosing the best one based on this history. The badness threshold parameter allows for maximizing cache capacity when the consistency level is ONE, ensuring availability even if a replica performs worse than others. Dynamic snitching doesn't determine replica placement itself; it simply wraps your chosen snitch and provides adaptive behavior on reads. In upcoming releases of Cassandra, improvements to dynamic snitching include using statistically significant random samples for updates, considering factors like compaction status when choosing a node, and penalizing unresponsive nodes based on time elapsed since the last reply.

Feedback from the recent Cassandra Summit highlighted a need for introductory content tailored to those new to Apache Cassandra. In response, DataStax has released a free "Introduction to Apache Cassandra" white paper aimed at tech and IT professionals exploring Cassandra for the first time. The paper provides an overview of Cassandra, its strengths and performance capabilities, as well as top use cases. Other helpful resources are also available for those getting started with NoSQL and Cassandra.

Apache Cassandra is a distributed database system that encourages denormalization instead of joins to improve performance. However, this approach can make some simple patterns less elegant. To address this issue, Cassandra 1.2 introduces three new CQL3 types: map, set, and list. These collection types provide more powerful and natural alternatives for handling multiple email addresses or other similar scenarios. The operations allowed on these collections include setting values, adding elements, removing elements by index or value, and updating specific elements. Collection support is already available in the Cassandra trunk and will be part of the final release of CQL3 in Cassandra 1.2.