Linux Server Administration

A server can have ample capacity on a data mount while the root filesystem fills with logs, package cache, container layers, or temporary files. Applications may still fail because key paths reside on root. What to check Check usage by mount and directory, inode counts, container storage, journal size, package cache, and open deleted files. Identify...

A locale change can affect date parsing, sorting, decimal separators, character handling, and command output. Scripts may behave differently even when the same files and software are present. What to check Check service environment variables, system and user locale, application logs, date and number parsing, and whether scheduled jobs inherit the same...

A service manager may restart a process after each crash or failed health check. The repeated restarts can consume resources and hide the original failure if logs rotate quickly. What to check Check service manager restart policy, exit code, recent logs, resource limits, dependencies, configuration, and crash timestamps. Capture one failure cycle and...

Package updates can be slow because of repository reachability, DNS, proxy configuration, signature verification, disk contention, or a repository mirror issue. The server may still have normal application response times. What to check Check package manager logs, repository URLs, DNS, proxy settings, mirror latency, disk use, and certificate or key...

Services can disagree about event times when the operating system, application, database, or user profile applies different timezone rules. Daylight-saving transitions can create missing or repeated local times. What to check Check system timezone, UTC clock, application settings, database storage format, log display rules, and scheduled task...

A service that works after a manual start may depend on a mounted volume, network route, environment variable, or another service that is unavailable during boot. What to check Check systemd status and journal entries around boot, unit ordering, permissions, environment files, mount readiness, and the service user’s working directory. Avoid repeatedly...

A filesystem can run out of inodes while still showing free storage space. Applications may then fail to create files, sessions, caches, or mail even though disk capacity looks available. What to check Check inode usage by mount, identify directories containing large numbers of small files, and confirm whether temporary files or application queues are...

A firewall rule can appear present but fail to allow traffic because of rule order, the wrong interface or address family, a separate cloud firewall, or a service listening only locally. What to check Check the listening socket, host firewall chain and counters, interface, source and destination addresses, IPv4 and IPv6 paths, and any upstream network...