Distributed database Apache Ignite 2.11 is now available and comes with partition awareness enabled, built-in snapshot restore and consistency check commands, and new page replacement policies. It’s also the first release to let teams use the RendezvousAffinityFunction affinity function with ClusterNodeAttributeColocatedBackupFilter to group nodes into virtual cells.
D2iQ boards the declarative train with DKP 2.0 release
D2iQ Kubernetes Platform 2.0 (DKP) has landed and marks the tool’s move towards a more declarative approach promising better automation throughout the infrastructure lifecycle.
According to D2iQ, this has been realised through a reworking of the platform’s architecture, which now makes use of the ClusterAPI project for auto-scaling resources as well as self-proclaimed GitOps tool FluxCD. Other than that DKP 2.0 comes fitted with capabilities for automatic node recovery, an improved Konvoy GUI, and Flatcar Linux support.
DeisLabs experiments with WebAssembly as a PaaS foundation
Cloud native experts DeisLabs have used their interest in the WebAssembly technology to come up with a new play on platform as a service. The result is called Hippo, an open source self-hosted PaaS project that promises an “easy way to scaffold and test out new ideas”. It also looks to simplify the management of the application release cycle through the concept of Channels that “automatically deploy the latest release based on the criteria provided”.
The use of WebAssembly under the hood is meant to make Hippo more secure than regular PaaS offerings, since all applications are sandboxed, portable and fast. And while it is still in its early stages, the team has big plans to expand capabilities and grow interaction with other systems, so Hippo could definitely at least be worth a look.
No internet access, no problem: JupyterLab makes its way onto the desktop
The team behind Jupyter UI JupyterLab this week announced the availability of a standalone application of its project called JupyterLab Desktop App. The software is based on Electron and bundles a Python environment with several popular Python libraries such as NumPy and pandas for easy use when there’s no internet access available. And to make it truly an option for everyone, the app is supposed to work on Debian and Fedora based Linux, macOS and Windows operating systems.
Google kicks cloud CD into higher gear
Teams who’ve had trouble integrating Google Kubernetes Engine into their Continuous Delivery process in the past and don’t want to give GitLab a try can now get help from the company itself. Google Cloud Deploy is a new managed CD service especially geared towards GKE. The still in preview offering promises users a central interface for managing releases through test, stage, and production environments without the need for manual infrastructure set-up. Teams are able to define pipelines declaratively, while the platform looks to provide best practices to ensure a setup’s speed and safety.
Of course Google made sure Cloud Deploy plays well with other company products such as IAM and Cloud Audit Logs, but it also assures easy integration with other standard tools from the DevOps space. Since the product is in its early stages it is free for now, once generally available customers will be charged per active pipeline each month (“no charge for the first active delivery pipeline per billing account each month, with each additional delivery pipeline charged at $15.00 per month”).