Episode 105 of The Linux Podcast with Fexingo dives into Kernel Same-page Merging (KSM), a memory deduplication technique that saves significant RAM in virtualized and containerized environments. Lucas and Luna explain how KSM scans memory pages, identifies identical content, and merges them into shared read-only pages, reducing the memory footprint of multiple virtual machines running similar operating systems or workloads. They discuss the trade-offs: the CPU cost of scanning, the potential for performance degradation under heavy write loads, and how to tune KSM parameters like pages_to_scan and sleep_millisecs. Real-world examples include cloud providers using KSM to overcommit memory in production. The episode also covers alternatives like zswap and the newer KSM improvements in recent kernels.

#Linux #KSM #MemoryManagement #Virtualization #Kernel #OpenSource #Technology #FexingoBusiness #BusinessPodcast #SystemAdministration #CloudComputing #Containers #Deduplication #RAM #KVM #QEMU #PerformanceTuning #LinuxKernel

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna dive into Linux kernel 6.12's landmark change: the BPF-based scheduler (sched_ext) replacing the classic Completely Fair Scheduler. They unpack how BPF allows scheduling policies to be written as safe, loadable programs — enabling experiments like Google's scx_layered for latency-sensitive workloads, and Meta's scx_rustland in production. The episode explains why this matters for cloud gaming, VM latency, and custom scheduling without kernel patches. With real examples from Meta's datacenters and the SCX project's open-source community, they show how this shift makes Linux scheduling extensible for the first time.

#Linux #Kernel #BPF #Scheduler #CFS #sched_ext #eBPF #Kernel6.12 #OpenSource #SCX #Meta #Google #CloudGaming #Latency #Performance #Technology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Linux Podcast with Fexingo, Lucas and Luna dive into Linux namespaces, the kernel feature that forms the backbone of container isolation. They explain how namespaces give each container its own view of the system—separate process trees, network stacks, mount points, and more—without the overhead of a full virtual machine. Lucas breaks down the seven namespace types (PID, net, mnt, uts, ipc, user, cgroup) and how they work together to create the illusion of a standalone system for each container. Luna shares a real-world example from a production incident where misconfigured PID namespaces caused a cascade of zombie processes. They also discuss how user namespaces add an extra layer of security by allowing unprivileged users to run containers safely. The episode ends with a practical tip: check your namespace configuration with 'lsns' and 'unshare' to understand what your containers actually see.

#Linux #Namespaces #Containers #ContainerIsolation #PIDNamespace #NetNamespace #MntNamespace #UtsNamespace #IpcNamespace #UserNamespace #CgroupNamespace #Docker #Kubernetes #OpenSource #Technology #Sysadmin #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Most Linux security conversations focus on external attacks—firewalls, SELinux, AppArmor. But what about the threat already inside your network? Episode 102 of The Linux Podcast digs into auditd, the Linux Audit Framework that's been quietly logging every system call, file access, and privilege escalation since kernel 2.6. Lucas walks through a real-world scenario: a developer with legitimate sudo access who accidentally exfiltrates customer data via rsync. He explains how audit rules catch the anomaly—rules that trigger on file reads to /etc/shadow, outbound connections from unexpected processes, and changes to cron jobs. Luna pushes on the operational pain: auditd logs can grow at gigabytes per hour on busy servers, and interpreting ausearch output is a skill in itself. They discuss tuning strategies, the golden ratio of rule specificity vs. performance, and why 'trust but verify' is the only sane posture in a zero-trust world. No hot takes. Just the mechanics of watching the watchers.

#LinuxAuditd #SystemCallAuditing #InsiderThreat #SecurityMonitoring #LinuxSecurity #AuditFramework #ZeroTrust #KernelAuditing #AccessControl #DataExfiltration #Compliance #SOC #ThreatDetection #DevSecOps #Technology #FexingoBusiness #BusinessPodcast #LinuxPodcast

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna dive into the Linux Kernel Crypto API, the subsystem that offloads encryption, decryption, and hashing to hardware accelerators like Intel QAT, AMD CCP, and ARM CryptoCell. They unpack how the API abstracts vendor-specific drivers behind a unified interface, enabling applications like dm-crypt (LUKS disks), IPsec, and TLS termination in the kernel to transparently use hardware crypto without user-space changes. The episode focuses on a specific example: how switching from software AES to hardware-accelerated AES on an Intel Xeon with QAT can reduce CPU load by over 80 percent while maintaining wire-speed throughput. Lucas explains the async request model, the role of cryptographic operation vectors (crypto_async_request), and how the crypto manager handles fallback to software if hardware is unavailable. Luna asks about real-world deployment gotchas — like driver maturity, NUMA affinity, and the trade-off between latency and throughput in synchronous vs. asynchronous paths. They also touch on how the same API powers kernel TLS (KTLS) for Netflix-scale content delivery. No marketing fluff — just a clear, technical look at a subsystem that quietly improves performance and security across millions of Linux servers.

#LinuxKernelCryptoAPI #HardwareAcceleration #IntelQAT #AMDCCP #ARMCryptoCell #AES #dm-crypt #LUKS #IPsec #KTLS #KernelTLS #CryptoOffload #LinuxSecurity #KernelSubsystem #Technology #LinuxPodcast #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In this 100th episode, Lucas and Luna explore Kernel Same-page Merging (KSM), a memory-saving feature in Linux that deduplicates identical memory pages across virtual machines and applications. They break down how KSM works under the hood—scanning pages, merging them into a single copy-on-write page, and managing the trade-offs with CPU overhead. The hosts discuss real-world use cases like running hundreds of VMs on a single host, KSM's tuning parameters like pages_to_scan and sleep_millisecs, and its evolution in recent kernel versions. They also touch on KSM's cousin, UKSM, and alternatives like transparent hugepage defragmentation. A practical episode for anyone managing memory-constrained Linux systems.

#Linux #KSM #KernelSamePageMerging #MemoryManagement #Virtualization #KVM #QEMU #OpenSource #LinuxKernel #Deduplication #MemorySaving #Hypervisor #TechDeepDive #FexingoBusiness #Technology #BusinessPodcast #LinuxPodcast #SystemAdministration

Keep every episode free: buymeacoffee.com/fexingo

Luke and Luna crack open the system call layer to explain how tools like strace, seccomp, and Falco intercept kernel calls for debugging, security, and container sandboxing. Using real examples—from tracing a Python web server's reads and writes to how Docker uses seccomp profiles by default—they demystify the mechanism behind system call interposition. They also discuss performance tradeoffs and why modern Linux replaced ptrace-based tracing with BPF-based solutions like the tracepoint and kprobe infrastructure. Perfect for developers who want to understand what happens between the application and the kernel, and how to safely observe or restrict it.

#Linux #SystemCalls #strace #seccomp #BPF #KernelTracing #ContainerSecurity #Falco #Docker #LWN #Technology #FexingoBusiness #BusinessPodcast #LinuxKernel #SyscallInterposition #Performance #Tracepoints #kprobes

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna dive into netfilter, the kernel framework that powers Linux firewalls — from iptables to nftables. They walk through the five Netfilter hooks that intercept packets at key points in the network stack, explain how connection tracking reassembles flows, and explore why modern tools like nftables and eBPF are replacing legacy iptables. Along the way they break down a concrete example: an incoming TCP SYN packet traversing the PREROUTING, FORWARD, and POSTROUTING hooks. If you have ever wondered how iptables rules actually get applied, this episode gives you the mental model.

#Netfilter #LinuxFirewall #iptables #nftables #eBPF #KernelNetworking #PacketFiltering #ConnectionTracking #NFtablesMigration #LinuxSecurity #NetworkStack #Technology #OpenSource #FexingoBusiness #BusinessPodcast #LinuxPodcast #PodcastShow #TechDeepDive

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Linux Podcast with Fexingo, Lucas and Luna dive into Kernel Same-page Merging (KSM) — a memory-saving technique that deduplicates identical pages across virtual machines and containers. They break down how KSM works, its performance trade-offs, and real-world use cases like large-scale server consolidation. Using concrete examples like a hosting provider running 200 VMs on a single host, they explain how KSM can reduce memory usage by 30-50% without sacrificing stability. Along the way, they touch on the merge_across_nodes debate, the role of ksmd and its CPU overhead, and when to use madvise vs. KSM. A practical, specific look at a feature that quietly powers cloud infrastructure everywhere.

#KSM #KernelSamePageMerging #Linux #MemoryManagement #Virtualization #KVM #Containers #CloudInfrastructure #ServerConsolidation #ksmd #MemoryDeduplication #OS #Technology #LinuxPodcast #FexingoBusiness #BusinessPodcast #Podcast #Episode97

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Linux Podcast with Fexingo, Lucas and Luna dive into the evolution of Logical Volume Manager (LVM) in Linux, focusing on a specific concrete case: how LVM thin provisioning and snapshots are being used by a mid-size SaaS company to cut storage costs by 40% while maintaining production performance. They break down the difference between traditional partitions and logical volumes, explain thin provisioning with real numbers, walk through a snapshot-based backup strategy that avoids data duplication, and explore LVM cache for SSD tiering. Lucas shares a hands-on example of resizing a live filesystem without downtime, and Luna challenges the complexity trade-off. This episode is perfect for sysadmins, DevOps engineers, and anyone managing server storage who wants practical, actionable insights into modern LVM features.

#Linux #LVM #StorageManagement #ThinProvisioning #Snapshots #LogicalVolumeManager #Sysadmin #DevOps #ServerStorage #SSD #Caching #Filesystem #Backup #DataRecovery #Technology #OpenSource #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo