Debian 12.4 has been released, superseding Debian 12.3 which had a bug that could potentially cause data corruption. The bug, which was reported under the bug advisory #1057843, concerned issues with kernel-image-6.1.0-14 (6.1.64-1). The latest release, Debian 12.4, includes fixes for this bug, along with other important bug fixes.

Debian 12.4 is an update to the stable distribution Debian 12, codenamed “bookworm”. This point release focuses on correcting security issues and addressing other serious problems. It is important to note that Debian 12.4 does not represent a new version of Debian 12, but rather updates certain packages included in the distribution. Users do not need to discard their old Debian 12 media, as they can simply upgrade their packages to the current versions using an up-to-date Debian mirror.

For users who regularly install updates from security.debian.org, there will be minimal package updates with this point release, as most of the updates have already been included. New installation images will be made available soon at the usual locations.

The update includes a comprehensive list of bug fixes for various packages. The complete list of bug fixes can be found in the Debian 12.4 Changelog. In addition to bug fixes, Debian 12.4 also includes security updates. The Security Team has released advisories for each of these updates, addressing vulnerabilities in packages such as Chromium, Firefox ESR, Exim4, Thunderbird, and more. The installer has also been updated to include the fixes incorporated into the stable release by the point release.

For more information about Debian 12.4, including the complete list of packages that have changed, the current stable distribution, proposed updates, and security announcements, visit the Debian website.

ASRock Industrial has released the 4×4 BOX-7040U series mini PC, featuring AMD’s Ryzen 7040U “Phoenix” architecture. The Ryzen 7040U chips combine Zen 4 CPU cores with RDNA 3 graphics, making them popular choices for handheld gaming PCs and laptops. These chips have now made their way into mini PCs, with Gigabyte and ASRock Industrial introducing their own models earlier this year. ASRock has also introduced two 4X4 motherboards, the 4×4-7640U and 4×4-7840U, for users who prefer to supply their own chassis.

The versions of the mini PC measure 118 x 110 x 48mm and offer support for up to 64GB of dual-channel DDR5-5600 memory. They also support an M.2 2280 PCIe Gen 4 SSD and/or a 2.5 inch SATA hard drive or SSD. The mini PCs come with WiFi 6E, Bluetooth 5.2, and support for up to four displays. The ports include 2 x HDMI 1.4b, 2 x USB4 (with DisplayPort 1.4a Alt Mode), 1 x USB 3.2 Gen 2 Type-A, 2 x USB 2.0 Type-A, 1 x 2.5 GbE Ethernet (RTL8125BG), 1 x Gigabit Ethernet (RTL8111EPV), and 1 x 3.5mm audio.

The 4×4 BOX-7840U features an AMD Ryzen 7 7840U chip with 8 CPU cores, 16 threads, and 12-core graphics. The 4×4 BOX-7640U has a Ryzen 5 7640U chip with a 6-core, 12-thread CPU and 8-core graphics.

Source: Liliputing.

One of the most notable features of the new Raspberry Pi platform is its small, vertical, 16-way FFC (Flat Flexible Cable) connector on the left-hand side of the board. This connector exposes a single-lane PCI Express interface.

The Peripheral Component Interconnect Express (PCI Express or PCIe) is a board-level interconnect that enables high-speed data transfer between a processor chip and external peripherals such as NVMe SSDs, Ethernet cards, and AI/ML accelerators. PCIe achieves this by serializing data transfers and sending one bit at a time down a single channel. Each channel consists of one or more differential pairs on the PCB, which are controlled waveguides made by closely spaced wires. In the case of a single-lane PCIe interface, there is a single transmit pair, a single receive pair, and a clock pair, requiring three differential pairs and six wires. The Raspberry Pi 5, based on the BCM2712 processor, is connected to the RP1 I/O controller via an ×4 interface.

The PCIe specification also requires sideband signals such as reset, clock request, and wakeup. The 16-way connector on the Raspberry Pi provides all these signals, along with two pins for controlling board power and automatically detecting a properly designed PIP (PCIe Peripheral) by the Raspberry Pi firmware.

Instead of adding an M.2 connector to the Raspberry Pi 5, which would be large, relatively expensive, and require a 3.3V, 3A power supply, the Raspberry Pi team opted for a small, low-cost FFC connector. This allowed them to provide a PCIe interface without increasing the board size or imposing additional costs on users.

At the launch of the Raspberry Pi 5, the team did not have a specification for building peripherals that attach to the 16-way PCIe connector. They wanted to thoroughly test their own prototype product and consider the interaction of PCIe peripherals with Raspberry Pi power states and firmware. They have now released the first revision of the specification and are in the final stage of prototyping their own M.2 M Key HAT+. This HAT+ will be launched early next year.

In addition to the PCIe connector specification, the Raspberry Pi team has also released a preliminary version of the new HAT+ specification. The original HAT specification, written in 2014, is in need of an update. The new specification simplifies certain aspects, including the required EEPROM contents, and consolidates everything into one document. It also adds new features. While there is still work to be done on this standard and the EEPROM utilities have not been updated, this release provides a glimpse into the changes in the HAT standard.

The Raspberry Pi team wanted to ensure that the HAT+ standard is developed correctly, as it is expected to be in use for as long as the old HAT standard. They believe that PCIe boards (PIPs) that go on top of the Raspberry Pi should be HAT+ boards, and their own M.2 HAT+ will adhere to this standard.

More information can be found in the first revision of the Raspberry Pi Connector for PCIe datasheet, and the Raspberry Pi HAT+ Specification datasheet.

Uptime Kuma, the self-hosted uptime monitor, has released version 1.23.9, bringing several improvements, bug fixes, and security fixes to the platform.

One important note is that this release may be a breaking change for those using third-party frontends or tools. The WebSocket origin now needs to be the same as your server hostname. However, users can set an environment variable called UPTIME_KUMA_WS_ORIGIN_CHECK to bypass in order to skip this check.

Here are the improvements included in this release:

• Added an aria-label to the monitor search box, improving accessibility.
• Added a helptext for the ntfy’s priority field, providing better guidance to users.

The bug fixes in this release are as follows:

• Corrected the Maintenance Start/End Time Input to Use Explicitly Specified Timezone, ensuring accurate time tracking.
• Fixed the buttons of ActionsSelect and ActionsInput that had a default type="submit", preventing unintended form submission.

In terms of security fixes, the following updates were made:

• Changing the password now closes all logged-in socket connections immediately, preventing unauthorized access.
• The WebSocket server can now only be connected from the same origin, similar to the CORS policy.
• An environment variable called UPTIME_KUMA_WS_ORIGIN_CHECK has been added, with two options: cors-like (default) and bypass.

Additionally, this release includes other small changes, code refactoring, and comment/documentation updates.

Netdata, a popular monitoring system, has released version v1.44.0 with several exciting new features and improvements. This release further solidifies Netdata’s position as a leading monitoring solution for servers, Linux, DevOps, and home labs.

One of the major highlights of this release is the significant improvement in performance, surpassing even Prometheus, a well-known monitoring system. Netdata now includes a new streaming protocol called SLOTS, which allows for more efficient metric streaming between children and parents. This reduces overhead on parents by about 30% without impacting the children. Additionally, Netdata now supports multiple compression algorithms, including ZSTD, GZIP, and BROTLI, with ZSTD being the default choice for its balance between compression ratio and CPU consumption.

Another major addition is the introduction of Gorilla compression, a time series data compression technique developed by Facebook for their time series database. When enabled, Gorilla compression provides a 30% reduction in memory usage for Netdata, making it even more efficient compared to Prometheus.

Netdata now also has improved support for handling large systemd-journal databases, making it more capable of dealing with huge log volumes. The systemd-journal.plugin has been optimized for performance in such environments, providing prompt responses to queries. Netdata’s logs have also been rewritten to log to the systemd-journal, allowing for easy monitoring and analysis using Netdata’s systemd-journal.plugin user interface.

A new utility called log2journal has been introduced in beta, allowing the conversion of log files into structured systemd-journal log entries. This powerful tool supports processing various log formats, including JSON and logfmt logs, and can be used to extract, convert, transform, and send logs to systemd-journal.

Netdata has also expanded its range of functions, offering new ways to visualize and troubleshoot system metrics. These functions leverage the wide range of collectors and metrics available in Netdata, providing insights into disk I/O activity, resource utilization of containers and virtual machines, IPMI sensor readings, disk usage for mount points, network traffic, process resource usage, and more.

In addition to these feature enhancements, Netdata has added new alert notification integrations to Netdata Cloud, including Amazon Simple Notification Service (Amazon SNS) and Telegram. These integrations provide users with more options for receiving alert notifications from Netdata.

It’s worth noting that some changes have been made in this release, including the removal of the charts.d/nut collector, which has been replaced by go.d/upsd. Netdata’s internal metrics are now disabled by default to reduce data volume, and Gorilla compression will be enabled by default in the next release. Some exporters, such as Google Cloud Pub Sub and AWS Kinesis, will be removed in the next release, and database modes map and save will also be eliminated. Furthermore, per-core CPU metrics will be disabled by default to improve performance, and several eBPF.plugin modules have been disabled to optimize system performance.

Overall, Netdata’s v1.44.0 release brings significant improvements in performance, log handling, compression, and functionality, making it an even more powerful and efficient monitoring system for servers, Linux, DevOps, and home labs.

For more details and to download the latest release, visit the Netdata GitHub page.

Harvester, an open-source hyperconverged infrastructure (HCI) solution built on Kubernetes, has released version v1.3.0-dev-20231208 for testing. Harvester is designed for operators who are looking for a cloud-native HCI solution and runs on bare metal servers. It offers integrated virtualization and distributed storage capabilities, supporting both traditional virtual machines (VMs) and containerized environments through integration with Rancher.

This release is specifically for testing purposes and comes with a few important notes:

• It is not fully tested, so users should proceed with caution.
• Upgrading from previous releases or to future releases is not supported.
• Bug reports are welcome.

Artifacts for this release include the following:

• harvester-v1.3.0-dev-20231208-amd64.iso
• harvester-v1.3.0-dev-20231208-vmlinuz-amd64
• harvester-v1.3.0-dev-20231208-initrd-amd64
• harvester-v1.3.0-dev-20231208-rootfs-amd64.squashfs
• harvester-v1.3.0-dev-20231208-amd64.sha512
• version.yaml

There have been several changes in this release, including bug fixes and feature updates. Notable changes include fixing the SupportBundle CRD additionalPrinterColumns path, restoring the start of VMs if they were voluntarily powered off, and updating various components such as Wharfie, KubeVirt, and Longhorn.

Overall, Harvester’s latest release provides users with an opportunity to test and explore its features and functionalities. However, it is important to remember that this release is not fully tested and should be used at one’s own risk.