In this episode, Lucas and Luna dig into Linux Kernel Samepage Merging (KSM) — the memory-saving technique that deduplicates identical pages across processes and containers. With KSM, a host running hundreds of duplicate container images can cut memory usage dramatically, letting you pack more workloads onto fewer servers. But KSM isn't a free lunch: it adds CPU overhead and can introduce security risks if not configured carefully. Lucas walks through how KSM works, from the kernel's scan-and-merge process to the userspace knobs like /sys/kernel/mm/ksm/. He cites a real example: a CI farm that reclaimed 30 percent of memory by enabling KSM on its build runners. Luna challenges the performance trade-offs and asks about the security side, especially the 'KSM security' concerns that have haunted it for years. They also touch on how KSM differs from other memory-saving tricks like transparent huge pages and memory cgroups. For anyone running containers or virtual machines at scale, this episode gives you a concrete tool to squeeze more out of your RAM. Plus, the hosts share a light-hearted moment about how KSM's name sounds like a friendly neighborhood initiative.

#Linux #KSM #KernelSamepageMerging #MemoryManagement #Containers #Virtualization #SystemAdministration #DevOps #ServerOptimization #OpenSource #Technology #Podcast #FexingoBusiness #BusinessPodcast #TechPodcast #LinuxKernel #MemoryDeduplication #CI

Keep every episode free: buymeacoffee.com/fexingo

Memory hotplug lets you add or remove RAM from a running Linux system without a reboot. It’s a niche but powerful capability used in cloud fleets, big databases, and high-availability setups. Lucas and Luna dig into the mechanism — from sysfs entries to the kernel’s memory block layer — and walk through a real-world example of resizing a VM’s memory on the fly. They also touch on the risks, the hardware requirements, and why most desktop users will never need it. If you’ve ever wondered how cloud providers seem to resize instances without downtime, this episode explains the trick.

#LinuxMemoryHotplug #RAMHotplug #LinuxKernel #MemoryManagement #CloudInfrastructure #VirtualMachines #Sysadmin #OpenSource #LinuxServer #HighAvailability #MemoryBlocks #Sysfs #Technology #LinuxPodcast #FexingoBusiness #BusinessPodcast #DevOps #InfrastructureAsCode

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Linux Podcast with Fexingo, Lucas and Luna dive into the Linux 'perf' tool, the built-in profiler that many sysadmins overlook. They explore how perf uses hardware counters and kernel tracepoints to pinpoint CPU bottlenecks, cache misses, and syscall stalls with minimal overhead. The conversation centers on a real-world scenario: a web server experiencing mysterious latency spikes. Lucas explains how to read a 'perf top' output, while Luna shares a tip about using 'perf record' and 'perf report' to spot a specific cache-miss pattern. They also compare perf to eBPF-based observability tools, clarifying when each is the right choice. By the end, listeners learn one concrete command they can run today—'perf stat'—and how to interpret its key metrics. If you've ever felt lost in a sea of monitoring dashboards, this episode gives you a direct path to understanding what your CPU is actually doing.

#Linux #PerfTool #PerformanceProfiling #CPUBottlenecks #CacheMisses #SyscallLatency #Observability #eBPF #ServerStack #Technology #FexingoBusiness #BusinessPodcast #OpenSource #Sysadmin #DevOps #Kernel #HardwareCounters #Tracepoints

Keep every episode free: buymeacoffee.com/fexingo

In this episode, Lucas and Luna dive into Btrfs snapshots, a feature that lets you roll back your system or files in seconds. They explain the copy-on-write magic behind snapshots, how to create and manage them with btrfs subvolume snapshot, and why they're a game-changer for home labs and production servers alike. Using the example of a botched system update that nearly wrecked a server, they show how a simple snapshot can be a lifesaver. They also cover snapshot management — automating them with snapper, handling disk space with btrfs balance, and the key difference between snapshots and traditional backups. If you've ever wished you could undo a system change, this episode is for you. Plus, they touch on how snapshots integrate with tools like Timeshift and openSUSE's Snapper, making Btrfs a compelling choice for modern Linux setups. Tune in to learn how to protect your data with Btrfs snapshots and sleep easier at night.

#Btrfs #Snapshots #Linux #Filesystem #CopyOnWrite #Subvolume #Snapper #Timeshift #OpenSUSE #HomeLab #DataRecovery #SystemRollback #LinuxAdmin #Backup #Technology #LinuxPodcast #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In Episode 161, Lucas and Luna dive into Linux kernel memory reclaim — the behind-the-scenes mechanism that keeps servers from falling over when memory runs out. They explore how the kernel's reclaim algorithms decide which pages to evict, the role of cgroups and memory limits, and why tuning swappiness and vfs_cache_pressure matters for real-world workloads. Using a concrete example of a busy web server under memory pressure, they explain the difference between anonymous and file-backed pages, the function of kswapd and direct reclaim, and how modern kernels use memcg to reclaim memory per container. Lucas also walks through a practical scenario: a Redis cache on a memory-limited host, and what happens when the kernel starts reclaiming its pages. The episode closes with a forward-looking question about the future of reclaim with tiered memory systems. Perfect for Linux sysadmins, SREs, and anyone running containers in production.

#Linux #Kernel #MemoryReclaim #cgroups #Kswapd #Swappiness #ContainerMemory #Sysadmin #SRE #Redis #WebServer #MemoryPressure #OpenSource #Technology #FexingoBusiness #BusinessPodcast #LinuxPodcast #ServerStability

Keep every episode free: buymeacoffee.com/fexingo

In this episode, Lucas and Luna explore F2FS, the flash-friendly file system designed by Samsung for NAND storage. They explain why traditional file systems like ext4 struggle with SSDs, how F2FS uses a log-structured design to reduce write amplification, and the role of multi-head logging and adaptive algorithms in improving performance and longevity. They also discuss real-world adoption, from Android smartphones to enterprise SSDs, and weigh the trade-offs of using F2FS for your own projects. Perfect for developers, sysadmins, and anyone curious about the technology behind their devices.

#F2FS #LinuxFileSystem #FlashStorage #SSD #NAND #Samsung #OpenSource #Kernel #Technology #SysAdmin #DevOps #StorageOptimization #WriteAmplification #LogStructured #Android #FexingoBusiness #BusinessPodcast #Linux

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Linux Podcast with Fexingo, Lucas and Luna dig into Kernel Samepage Merging (KSM) — the Linux kernel feature that deduplicates identical memory pages across processes. They explain how KSM works, why it matters for virtualization and container workloads, and where it can actually hurt performance. Using the classic example of multiple virtual machines running the same guest OS, they break down the trade-offs between memory savings and CPU overhead. They also touch on practical tuning: when to enable KSM, how to monitor its effectiveness with tools like ksmctl, and why it's often disabled by default in modern hypervisors like KVM. This episode gives you a concrete mental model of memory deduplication in Linux, plus a few real-world considerations for your own servers.

#Linux #Kernel #KSM #MemoryManagement #Virtualization #KVM #Containers #Docker #SystemAdministration #DevOps #ServerStack #OpenSource #Technology #FexingoBusiness #BusinessPodcast #CloudComputing #MemoryDeduplication #LinuxPerformance

Keep every episode free: buymeacoffee.com/fexingo

In episode 158 of The Linux Podcast with Fexingo, Lucas and Luna dig into the real-world impact of combining io_uring with eBPF for high-speed networking. They unpack how these two kernel technologies work together to slash latency and boost throughput, using the example of a hypothetical CDN edge server handling millions of requests. The conversation covers the mechanics of both systems, the challenges of integration, and practical advice for developers looking to experiment. They also touch on the broader trend of kernel-bypass alternatives and why io_uring and eBPF are becoming the go-to for performance-critical applications. A concrete, accessible look at cutting-edge Linux internals.

#Linux #io_uring #eBPF #Networking #Kernel #Performance #Technology #OpenSource #CDN #XDP #SysAdmin #DevOps #Cloud #Server #FexingoBusiness #BusinessPodcast #TechPodcast #LinuxPodcast

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna dive into eBPF, the kernel technology that lets you trace network packets and system calls without adding any overhead. They start with a concrete example: using bpftrace to answer a question that would normally take hours of digging through tcpdump logs. Along the way, they explain what eBPF is, why it's a game-changer for observability, how it compares to older tools like tcpdump and strace, and why the Linux kernel can run these programs safely. They also touch on how companies like Netflix and Facebook use eBPF in production, and what that means for anyone running a server today. If you've ever wanted to see inside your kernel, this episode gives you the perfect entry point.

#eBPF #Linux #Observability #Kernel #Bpftrace #Networking #SysAdmin #DevOps #Performance #Trace #FexingoBusiness #Technology #OpenSource #LinuxKernel #CloudNative #Troubleshooting #PacketAnalysis #Security

Keep every episode free: buymeacoffee.com/fexingo

Episode 156 of Fexingo's Linux podcast dives into the rare but powerful pairing of io_uring and XDP for high-speed networking. Lucas and Luna explain how io_uring's asynchronous I/O model meets eXpress Data Path's early packet processing, cutting latency and CPU overhead in demanding server workloads. They use a concrete example from a typical cloud gateway, walk through the mechanics of both technologies, and discuss where the combination shines—and where it doesn't. If you've wondered how modern Linux handles millions of packets per second without melting the CPU, this episode gives you the mental model and the pitfalls to watch. No fluff, just the technical details that matter for sysadmins, developers, and anyone running network-heavy services.

#LinuxNetworking #io_uring #XDP #HighPerformanceComputing #ServerTech #NetworkStack #AsyncIO #Kernel #SysAdmin #DevOps #CloudInfrastructure #PacketProcessing #Technology #OpenSource #FexingoBusiness #BusinessPodcast #TechPodcast #LinuxPerformance

Keep every episode free: buymeacoffee.com/fexingo