Linux Server Admin with Fexingo: Sysadmin, Bash, and Server Engineering

Episode 29 of Linux Server Admin with Fexingo dives into the hidden bottleneck in systemd's journald: the default rate-limiting settings that silently drop log entries during bursts, making debugging impossible when you need logs most. Lucas and Luna walk through a real scenario where a busy web server was losing critical audit records because journald capped the rate at 10,000 messages per 30 sec...

In this episode of Linux Server Admin, Lucas and Luna dive into AppArmor—the mandatory access control system that's simpler than SELinux but powerful enough to confine major services like Nginx, MySQL, and Apache. They walk through a real-world case: a misbehaving PHP script that tried to write to /etc/passwd on a production web server, and how an AppArmor profile blocked it instantly. Lucas expla...

Lucas and Luna dive into a practical server engineering headache: timeouts. Lucas explains why the default systemd service timeout of 90 seconds can cause cascading failures in production, especially for long-running database migrations or backup jobs. He walks through creating drop-in units to override timeout settings without touching vendor-provided service files, using a real case from a Postg...

In this episode, Lucas and Luna dive into SELinux on Linux servers — specifically, why and how to implement targeted policies without breaking production workloads. Lucas walks through a real case from a financial services company that reduced their kernel-level vulnerability exposure by over 80 percent after enabling SELinux in enforcing mode on 200 servers. They discuss the common pain points: l...

Episode 25 of Linux Server Admin with Fexingo: Sysadmin, Bash, and Server Engineering dives into SSH key hygiene. Lucas and Luna walk through a real-world scenario: a startup that accumulated 400 authorized SSH keys across 60 servers, with keys tied to former employees and contractors still active. They explain why this happens, how to inventory keys using simple Bash scripts and tools like ssh-au...

Episode 24 of Linux Server Admin with Fexingo digs into a critical but often overlooked security gap: the sudoers file. Lucas and Luna walk through a real-world misconfiguration that gave an attacker passwordless root access on a monitoring server, and explain how to lock down sudo with defaults, timed credentials, command aliases, and the sudo_logsrvd audit daemon. They cover the new sudo 1.9.14...

Episode 23 of Linux Server Admin with Fexingo. Lucas and Luna dig into one of the most common yet misunderstood server failure modes: swap thrashing. They walk through a real scenario where a Postgres query triggered the OOM killer, and explain why the default vm.swappiness of 60 is almost always wrong for production. You'll learn exactly how to set swappiness to 10, configure systemd-oomd, and ad...

Lucas walks Luna through the blue-green deployment pattern, a strategy that eliminates application downtime during updates by running two identical production environments and switching traffic atomically. They break down the exact steps: provisioning a standby environment, running health checks on the new version, flipping the load balancer, and rolling back instantly if something breaks. Lucas u...

Episode 21 of Linux Server Admin with Fexingo dives into the overlooked art of outbound email from headless servers. Lucas and Luna unpack why cron output, system alerts, and automated reports often vanish into the void, and how a tiny tool called nullmailer fixes it without a full mail stack. They walk through a real-world case: a staging server silently dropping backup failure alerts for six mon...

In this episode, Lucas and Luna dive into why a dedicated WireGuard VPN mesh beats traditional VPN setups for server-to-server communication. They walk through a real-world example of a three-node cluster where tailscale-style WireGuard overlay simplified firewall rules, reduced latency by 12 milliseconds, and cut configuration time from hours to minutes. Lucas explains the kernel-level efficiency...

Before you run apt upgrade or yum update, there's a better safety net than hoping you can roll back. In this episode, Lucas and Luna dive into why filesystem snapshots — using LVM or ZFS — should be a standard pre-update ritual. They walk through a real scenario: a production server that survived a broken kernel update only because the sysadmin had a five-second snapshot routine. They explain the...

Episode 18 of Linux Server Admin with Fexingo tackles the Linux kernel's Out-Of-Memory (OOM) Killer—a necessary but brutal mechanism that can kill critical services when memory pressure spikes. Hosts Lucas and Luna explore why the OOM Killer often targets the wrong process, how systemd's OOMD (introduced in v243) gives sysadmins fine-grained control over which services get sacrificed, and a real-w...

Episode 17 of Linux Server Admin with Fexingo dives into the case for a dedicated log aggregation pipeline. Lucas and Luna walk through a real-world scenario where a production outage was prolonged by 45 minutes because logs were scattered across 12 servers with no central aggregation. They explain why tools like the ELK stack or Loki reduce mean time to resolution, how a centralized pipeline turn...

In this episode of Linux Server Admin, Lucas and Luna explore why a dedicated time-series database (TSDB) like InfluxDB or VictoriaMetrics can transform how you handle server metrics, logs, and monitoring data. They walk through a real-world scenario: a mid-size e-commerce company whose Prometheus setup buckled under 500,000 metrics per second, leading to dropped data and skewed alerts. Lucas expl...

In this episode of Linux Server Admin with Fexingo, Lucas and Luna argue that every production Linux server should run its own dedicated monitoring stack, separate from the application infrastructure. They walk through a real-world case where a shared Prometheus instance caused a cascading failure in a SaaS company's monitoring, and explain how a dedicated stack with Grafana, Prometheus, and Alert...

Episode 14 of Linux Server Admin with Fexingo: Sysadmin, Bash, and Server Engineering. Lucas and Luna explain why routing all SSH traffic through a dedicated bastion host dramatically reduces your attack surface and simplifies compliance. They walk through a real-world setup using a $5/month VPS as a jump box, discuss sshd hardening (disabling root login, key-only auth, port knocking), and cover a...

Lucas and Luna explain the concept of immutable root filesystems for Linux servers—why making the root partition read-only can stop configuration drift, prevent malware persistence, and simplify updates. They dive into how Fedora Silverblue and openSUSE MicroOS implement this, the trade-offs for package management, and why it's gaining traction for production workloads in 2026. No theoretical fluf...

Episode 12 of Linux Server Admin with Fexingo digs into a common but overlooked failure point in server management: running Ansible from a personal workstation. Lucas and Luna walk through a real-world case where a sysadmin's laptop battery died mid-playbook, leaving a fleet of 200 production servers in an inconsistent state. They explain why a dedicated, stateless Ansible control machine — a chea...

Episode 11 of Linux Server Admin tackles a silent server killer: clock drift. Lucas and Luna explain why even a 50-millisecond offset can break database replication, certificate validation, and distributed consensus. They walk through a real example of a PostgreSQL cluster that failed silently because chrony wasn't configured to poll multiple upstream servers. They compare ntpd vs chronyd vs syste...

When your production server suddenly drops into a kernel panic, the instinct is to reboot and hope it doesn't happen again. Lucas and Luna walk through a real case from a fintech startup where a misconfigured memory module caused intermittent panics across a 200-node cluster. They explain how to capture crash dumps, read the stack trace to identify the offending driver or hardware, and use tools l...

In this episode of Linux Server Admin with Fexingo, hosts Lucas and Luna explore the real-world impact of container orchestration on Linux server engineering. They trace the journey from early Docker days to Kubernetes becoming the default deployment platform for everything from microservices to batch jobs. Lucas breaks down why container orchestration isn't just for tech giants: small teams can s...

Episode 8 of Linux Server Admin with Fexingo. Lucas and Luna drill into disk I/O latency — the silent performance killer that most sysadmins ignore until it's too late. They walk through a real 2025 case: how a mid-sized e-commerce platform lost 12 percent of checkout conversions because their monitoring dashboard showed CPU and RAM as green while storage latency was spiking. You'll learn what 'io...

In this episode, Lucas and Luna dive into the practical world of server backups using Borgmatic, a wrapper for the Borg backup tool. They walk through a real disaster scenario where a rogue script deleted a week's worth of customer transactions on a staging server, and how a Borgmatic backup saved the day. The discussion covers the key advantages of deduplication, encryption, and incremental backu...

Lucas and Luna make the case that structured server documentation—not just comments in config files, but a living runbook—is the single most impactful investment a sysadmin can make. They walk through a concrete example: a senior engineer at a mid-size SaaS company who spent two weeks building a Markdown-based runbook repository, and how it saved the team 40 hours during a critical PagerDuty escal...

Memory leaks are the ghosts of server administration: invisible until your OOM killer reaps a production process at 3 a.m. In this episode, Lucas and Luna walk through a real case from a mid-sized ad-tech company where a 12-line Node.js script silently consumed 32 GB over six days. They explain how to spot the early signs using /proc/meminfo, why RSS vs. USS matters, and the one tool (valgrind's m...