Down a Rabbit Hole: The Importance of CI/CD
Open source developers working with Arm have to go down…
StorPool is distributed storage software which runs on standard hardware – servers, drives, network – and turns them into high-performance storage system. Its new 18.02 release includes support for arm64, block storage for Kubernetes, and Ubuntu 18.04 LTS.
Boyan Krosnov writes about the port to Arm and how current server designs stack up for the storage use case.
In particular use-cases, they can help us achieve unprecedented power efficiency and storage density. Especially Arm server chips with many PCIe lanes, are very interesting for high-density NVMe storage nodes and for high-density HDD storage nodes.
Interested in building an LED lighting project with Raspberry Pi? We’ve got you covered! In this post we’re going to take a look at setting up a custom RGB LED matrix and driving it with a Raspberry Pi, Fadecandy board and Glediator software.
Complete build details are included, and the system is completely self-contained and controlled from the Raspberry Pi. Balena deploys the necessary code via containers on the Pi, allowing the entire software stack to be deployed with a minimum of fuss.
FunFun Yoo from XSLAB shows photos of the V-Raptor, a 24-core Arm microserver. The design is based on the Socionext SoC with a BMC of their own manufacture. The target release date is February 2019.
Packet and Netronome are building microservers with Ampere eMAG CPUs. The design will fit in an Open19 rack and will host from four to eight servers in a single 1U form factor. The Netronome “smart NIC” allows developers to write code and deploy it via eBPF and execute on the network interface card instead of the CPU. The server design will also come in Intel and AMD flavors.
As announced last week, Firefox has announced support for its browser on Windows 10 for arm64. Now nightly builds are available for you to try the work in progress, report bugs, and contribute to Firefox development.
Numba is an open source JIT compiler that translates a subset of Python and NumPy code into fast machine code. In the latest 0.42 release candidates, it now has some support for 32-bit Arm (through the “Berryconda” project) and for 64-bit Arm (through the “Jetconda” project).
The first pieces of Rancher’s announced forthcoming support for arm64 and a lightweight Kubernetes distribution for the edge have emerged as a Hypriot based example for the Raspberry Pi 3.
The FreeBSD Foundation provided support to Alexandru Elisei, a student at University Politehnica of Bucharest, to port the bhyve hypervisor to arm64. Alexandru is now working at internship at Arm in Cambridge in the System Validation on Operating Systems team. His presentation on “Virtualization on ARMv8A bhyvearm64 Current Status” was given at Asia BSDCon in March 2018.
Julia is a scientific computing language that uses the LLVM back end to support code generation. The Julia 1.0.3 release now has support for arm64.
Weave Scope automatically generates a map of your application, enabling you to intuitively understand, monitor, and control your containerized, microservices-based application. Carlos Eduardo has undertaken a port of this package to arm64 and has a set of patches that are under review.
The release of Fedora 29 for the Raspberry Pi 3 has seen several kernel-level issues with support for wifi on that device. Peter Robinson of the Fedora Project report that it “should be fixed with kernel 4.19.10 for F-28/F-29.”
Roaring bitmaps are compressed bitmaps which tend to outperform conventional compressed bitmaps such as WAH, EWAH or Concise. They are used by several major systems such as Apache Lucene and derivative systems such as Solr and Elasticsearch, Metamarkets’ Druid, LinkedIn Pinot, Netflix Atlas, Apache Spark, OpenSearchServer, Cloud Torrent, Whoosh, InfluxDB, Pilosa, Bleve, Microsoft Visual Studio Team Services (VSTS), and eBay’s Apache Kylin.
A recent set of patches by Andrei Gudkov add support for vector operations (NEON or SIMD) for arm64, and result in substantial performance improvements (up to 5x for some microbenchmarks). These are available on the “master” branch at the moment and have not yet been released.
Darren Shepherd has put together a minimalist version of Kubernetes he calls “k3s” (or “kates”). The design goal is to pull out as many dependencies and features as possible while retaining the essential elements of Kubernetes so that apps continue to work.
The goal of the project is small simple (like swarm simple) k8s clusters. Use cases include small dev shops, CI/CD, edge, gpu, hobbyist, ARM toys