January 2017 Summaries

Brandon Lamb's blog post discusses his exploration of using ScyllaDB within LXC/LXD containers on colocated servers as part of a side project aimed at reducing cloud costs and enhancing database scalability and reliability. He initially considered using Cassandra but found ScyllaDB appealing due to its performance benefits. Lamb describes the setup process, including configuring the network using Ubuntu 16.04 on a Dell R610 machine, setting up LXC profiles for hugepages, and applying resource limits to prevent container resource contention. He details the steps taken to configure the environment, such as setting up networking with a bridge, creating and applying LXC profiles, and pulling an Ubuntu image for the containers. The post also covers installing ScyllaDB, configuring its server settings, and adjusting the `scylla.yaml` file to optimize resource usage within the containers. Lamb acknowledges missing some details and considers breaking down the process into multiple posts to provide more comprehensive guidance.

ScyllaDB 1.6 RC1 marks the release of the first Release Candidate for ScyllaDB 1.6, focusing on stability and bug fixes while introducing new features like experimental Debian support and upgradable Amazon Machine Images (AMIs). As an open-source, Apache-Cassandra-compatible NoSQL database known for superior performance and low latency, ScyllaDB 1.6 introduces changes to its metrics format, requiring users to update their monitoring dashboards accordingly. The release also addresses notable bug fixes, such as improving data streaming used in repairs, reducing bloom filter overhead, and resolving issues related to disk space and system performance under heavy load. Users are encouraged to report any bugs to help stabilize the release, with the note that only the last two minor releases will be supported once 1.6 is officially out.

ScyllaDB, a high-performance NoSQL database, offers the capabilities of Apache Cassandra with enhanced speed, allowing for simpler data models and reduced project complexity. By deploying ScyllaDB on Oracle Bare Metal Cloud Services, users can leverage the platform's robust CPU, memory, and storage resources to achieve maximum throughput and minimal latency. The deployment process involves configuring networking ports, creating instances, and setting up RAID for NVMe drives, followed by installing and configuring ScyllaDB. Benchmark tests demonstrated ScyllaDB's ability to handle large data volumes with high transaction rates, benefiting from Oracle's infrastructure to support efficient read and write operations. The cost for deploying ScyllaDB on Oracle's platform is estimated at $12.96 per hour for infrastructure alone, underscoring ScyllaDB's efficiency and performance potential in cloud environments.

In 2016, the ScyllaDB blog featured several popular posts that explored technical aspects of database management and optimization, including the asynchronous processing in networking versus storage, designing a userspace disk I/O scheduler for ScyllaDB, and tuning CPU schedulers to reduce latency spikes in ScyllaDB's 1.0.x version. The blog also discussed the challenges of maintaining predictable low latencies and the necessity of troubleshooting when encountering unexpected high latencies, as well as addressing common questions about determining optimal write request rates and resource allocation for NoSQL solutions. The content aimed to guide users in improving their ScyllaDB deployments and solicited feedback for future topics.