Grafana and Alertmanager. For graphing and sending webhooks into Discord and Email for me to see in real-time.
Prometheus and its exporters. If it has an exporter I can monitor it in Grafana. If it doesn’t have an exporter, I might be able to make one for it, since it’s just a plaintext static webpage with serialized metrics on it, anything can output that like PHP.
Infping/Influxdb for smokeping-like data in Grafana, works great for telling me about any packet loss or increased latency on any network that responds to ICMP echo. My uncle uses my services to complain to Spectrum about his Coax connection, which has led to him being able to have his ISP physically come out and stabilize it, now it hardly ever blips.
Also all of this is federated so if one server goes down I don’t lose 100% of my monitoring, and I can customize my data retention time based on how much disk I have locally. I use Ansible to automate updates and configuration.
It costs me 2-3 small-to-medium sized VPSes a month, or under $30/mo USD, which is a bargon compared to enterprise cloud solutions/services you can pay for that have small data retention or other limits. New Relic as an example is expensive if you want anything more than 30 days of data and they don’t offer as much as I get with my own monitors.
I’ve been running this setup since 2015 or so, I’ve used this setup to improve my craft and professional skills. I work with Linux Servers every day.
Because Bluetooth is a separate hardware module than the CPU. “Sleep” is just a low-power state for the CPU, one of the “S” states. Other modules on the motherboard are still powered and can handle their own tasks, like Wake on LAN received at your network card, or keeping your RAM hot with your running programs.
Everyone I know uses seatbelts. I’ve seen defeat devices that plug into the seatbelt receptacle but I don’t know anyone who uses one. Every vehicle I’ve been in has seatbelts except school buses and some charter buses.
To be honest, Ubuntu likely has nothing to do with it and I find the headline therefore misleading. It’s mostly the Linux kernel from how it reads.
Ubuntu 23.10 was run for providing a clean, out-of-the-box look at this common desktop/workstation Linux distribution. Benchmarks of other Linux distributions will come in time in follow-up Phoronix articles. But for the most part the Ubuntu 23.10 performance should be largely similar to that of other modern Linux distributions with the exception of Intel’s Clear Linux that takes things to the extreme or those doing non-default tinkering to their Linux installations.
I’ve been utilizing Fedora as my go-to server operating system for over a decade, starting around Fedora 13 and consistently progressing through the subsequent major releases. Throughout this journey, Fedora has proven to be a reliable choice, offering several noteworthy advantages.
One of the standout features of Fedora, reminiscent of its desktop counterpart, is the availability of up-to-date and cutting-edge packages. With the backing of Red Hat, compatibility with a range of software, including SystemD and FirewallD, enhances its appeal for server applications.
However, it’s crucial to consider Fedora’s rapid release cycle, with a new version emerging approximately every six months and a 13-month support window for each release. While this frequent update schedule ensures access to the latest features, it can pose challenges for server environments where uptime is critical, and system administrators may find it demanding to keep up with the pace while managing compliance, audits, and other business processes.
Despite these considerations, my personal experience with Fedora as a server has been exceptionally positive. The OS has demonstrated robust performance across diverse environments, seamlessly adapting to various setups, from bare metal and virtual machines to containers. The flexibility extends from smaller hardware configurations like the Raspberry Pi to more substantial servers with 40+ vCPUs and 1.5TB RAM.
SELinux, a crucial component for security, seamlessly integrates with most applications, but I have encountered some challenges with WINE headless server programs, particularly for hosting Windows-based game servers not originally designed for Linux. Additionally, compatibility with Debian-based Crypto wallets proved to be a stumbling block, requiring the use of a Debian VM for compilation and subsequent transfer to Fedora.
In terms of server redundancy and monitoring, I rely on three Fedora servers equipped with essential software such as Nginx, Grafana, Prometheus, Influxdb, and fping. Wireguard and Samba facilitate automated file transfers between another set of servers, contributing to an efficient setup.
Automation plays a significant role in my server management, with Ansible and Bash scripts streamlining tasks such as package installations, configuration adjustments, and firewall rule setups. Each Fedora server is configured for self-startup to ensure uninterrupted operation in case of power failure or manual reboots.
I’ve successfully integrated an LSI/Intel server RAID card into one of my Fedora servers, and the third-party RPM package for monitoring (storcli64) installed seamlessly. LSI card driver support is built into the Linux kernel, contributing to a hassle-free experience.
The only notable limitation I’ve encountered pertains to desktop gaming, a domain where I’m exploring transitioning to Fedora. For gaming servers requiring Steam/SteamCMD and Windows DLLs, alternative solutions might be more suitable, unless willing to navigate potential challenges with WINE and SELinux adjustments.
In conclusion, the suitability of Fedora as a server hinges on your specific use case. For web or Internet of Things servers, it stands out as an excellent choice. However, if your requirements involve gaming servers with Steam dependencies, alternative options may warrant consideration. As I’ve explored various Linux and *BSD systems over the years, Fedora has proven its versatility, offering a compelling solution for a range of server applications.
Fedora uses /var/www. Dunno what gonk you read or told you otherwise. There’s SELinux policies built in for that directory. You probably are confusing the default html files at /usr/share/html. These are separated intentionally. The /usr/share/html directory is managed by RPM, the other /var/www is content designated as web server files.
3·7 months ago
3·9 months ago
2·10 months ago
deleted by creator