May 2018 Summaries

In the blog post, the concept of Materialized Views in ScyllaDB is explored as a means to enhance the efficiency of data management within the Mutant Monitoring System. Materialized Views enable applications to maintain multiple data views without complex application logic, as they automate updates to additional view tables whenever changes occur in the base table. This feature simplifies data handling by shifting the complexity from the application level to the server level, allowing for faster sorting and storing of data. Despite their potential, Materialized Views are still experimental in ScyllaDB versions 2.1, 2.2, and 2.3. The post provides a step-by-step guide on setting up a ScyllaDB cluster, using a load generator tool, and creating a Materialized View that retrieves specific data columns, illustrating the practical application of this feature in enhancing data analysis capabilities.

ScyllaDB Enterprise 2018.1.2 is a minor release aimed at fixing bugs in the ScyllaDB Enterprise 2018.1 branch, which is based on the open-source ScyllaDB version 2.1, with additional bug fixes from earlier releases and enterprise-specific issues. Notably, this release resolves a reported issue with the IBM POWER8 architecture, allowing users to install ScyllaDB Enterprise 2018.1 on RHEL for IBM POWER8 systems. The release also addresses problems that could cause core dumps when closing secure connections or dealing with commit log errors. ScyllaDB Enterprise customers are encouraged to upgrade in coordination with the support team and explore further ScyllaDB resources, including installation options for AWS, local virtual machines, or Docker environments.

ScyllaDB Enterprise 2018.1.1 is a minor bug fix release for the 2018.1 branch of the enterprise NoSQL solution, derived from ScyllaDB open source 2.1 and incorporating bug fixes from previous versions as well as exclusive enterprise fixes. This release addresses several issues, such as problems with RAID 0 data directory access when upgrading the RHEL kernel, schema update redundancies during upgrades from ScyllaDB 2017.1 to 2018.1, and multi-data center write failures during schema changes. Additionally, the release specifies a dependency on newer kernels beyond kernel-3.10.0-514.el7, restricting installations on older kernels. Customers are advised to upgrade with support team coordination, and those using IBM POWER8 architecture are recommended to wait for further updates before upgrading.

Ola Cabs, a leading ride-hailing service in India, adopted ScyllaDB as a database solution to meet the demands of its real-time business, which often experiences quick spikes in demand due to various factors like weather and special events. Initially using Cassandra, Ola Cabs sought a database that could provide high throughput, low latency, and high availability, without the latency issues associated with the JVM runtime. ScyllaDB, written in C++, offered a promising alternative with its close-to-the-hardware approach and assured development of important features. Over a two-year period, Ola Cabs successfully integrated ScyllaDB into its ecosystem, using it for both write and read operations, and achieving improved performance by utilizing NVMe disks and configuring appropriate replication factors across multiple availability zones. The company found that using ScyllaDB without a caching layer simplified their architecture and reduced costs, while native metrics monitoring with Prometheus facilitated effective debugging and performance optimization. Ola Cabs continues to engage with ScyllaDB's community and anticipates future enhancements like secondary indexes and user-defined types to further support their mission-critical operations.

Phillip Tribble's blog post, part of a series on ScyllaDB training, discusses optimizing a Java application for the Mutant Monitoring System using prepared statements to improve scalability and performance. The series has undergone updates, available on ScyllaDB University, and this installment focuses on enhancing the Java application initially built with basic CQL statements. Prepared statements, which allow queries to be parsed and saved for reuse, help developers modify variables in queries efficiently, thereby boosting application performance. The post provides a step-by-step guide to incorporating prepared statements into the application, including adding necessary libraries and modifying functions to handle insert and delete operations with bound variables. The author demonstrates the process of building and running the updated application in a Docker environment, emphasizing the importance of prepared statements for efficient application development. The blog encourages continued exploration of prepared statements for enhancing application efficiency and provides resources for further learning about ScyllaDB.

Phillip Tribble's blog post details the process of creating a Java application that interacts with a ScyllaDB cluster, utilizing the Datastax Cassandra driver for Java. As part of the Mutant Monitoring System series, this installment focuses on connecting to the cluster, displaying, inserting, and deleting data from the Mutant Catalog table, and running the application within a Docker container. The post provides a step-by-step guide on setting up the environment, including importing necessary libraries and configuring Docker to run the application efficiently. It also hints at future topics, such as exploring prepared statements with the Java driver, while encouraging developers to enhance their understanding of ScyllaDB for monitoring purposes.

The ScyllaDB team has released version 2.0.4, a bugfix update for the ScyllaDB 2.0 stable branch, which maintains backward compatibility and supports rolling upgrades. Users are encouraged to upgrade to the latest stable open-source version, ScyllaDB 2.1.3. This release addresses several bugs, including issues with streaming failures, CQL prepared statement evictions, authentication request caching, and errors related to dropping keyspaces with User Defined Types. Additionally, it resolves problems such as ScyllaDB aborting due to partition queries, segmentation faults triggered by gossip bugs, and TLS session closure errors. Tools like the sstableloader, which previously struggled with specific data types, have also been improved in this update.

ScyllaDB 2.1.3 has been released as a bugfix update to the 2.1 stable branch, ensuring backward compatibility and support for rolling upgrades, as with all 2.x.y releases. The release addresses specific bugs, including an error encountered when dropping a keyspace with a User Defined Type (UDT) and a lack of support for UDT in the SSTable upload tool, scylla-tools-java. Users are encouraged to report any issues they encounter, and resources are available for obtaining ScyllaDB 2.1.3 via Docker, binary packages, and EC2 AMI, as well as guidance for upgrading from earlier 2.1.x versions.

In the tenth installment of the Mutant Monitoring System blog series, the focus is on teaching ScyllaDB administrators how to effectively back up and restore data within the Mutant Monitoring System, following a new policy by Division 3 in response to recent violent events. The process includes starting up the ScyllaDB Cluster, backing up the schema, and creating snapshots of key data spaces using nodetool. A simulated data loss scenario is conducted by deleting keyspaces to prepare for potential cyberattacks, after which the restoration process is demonstrated by recreating keyspaces from backups, restoring snapshots, and ensuring data integrity through repair commands. The blog underscores the importance of backups and provides insight into the technical steps necessary for maintaining data resilience within ScyllaDB environments.

Intel's Memory Group, known for developing the high-performance Optane SSD, partnered with ScyllaDB to create a big memory system capable of operating at high-volume scales. They selected ScyllaDB for its ability to fully utilize hardware capabilities and deliver optimal performance, noting its compatibility with Apache Cassandra that required no changes to existing applications. The group valued ScyllaDB's capability to optimize workloads without manual intervention or downtime, appreciating its open-source development approach that ensures rapid delivery of solutions. Frank Ober, a Solutions Architect at Intel, shared his positive experiences with ScyllaDB, highlighting its advantages in a video.

ScyllaDB Enterprise 2018.1.0 is a major release that enhances the production-ready ScyllaDB Enterprise, offering unique features, bug fixes, and improvements over the open-source variant. Key features include a new auditing capability for tracking user activity across multiple deployments, IBM Power support, improved performance for CPU-bound workloads, and a Heat Weighted Load Balancing system to optimize node selection. The release introduces Time Window Compaction Strategy for better handling of time series data, support for new data types and CQL improvements, and an updated intra-node sharding algorithm for enhanced data distribution among cores. ScyllaDB Enterprise 2018.1 requires Linux kernel 3.10.0-514 or later and supports ext4 file systems in production. Upgrades are possible from ScyllaDB Enterprise 2017.1.6 or ScyllaDB Open Source 2.1, with a recommendation to update Cassandra drivers for compatibility. The release also addresses known issues and provides guidance on installation, upgrades, and further learning resources.