Linux Server Administration

A filesystem can fill rapidly from logs, temporary files, deleted-but-open files, caches, or an application writing unexpected data. Checks to make Check filesystem and inode usage, identify largest directories, inspect open deleted files, and correlate growth with process and log timestamps. Preserve needed evidence and avoid deleting files blindly...

A daily spike commonly aligns with cron jobs, log rotation, report generation, backup, or traffic patterns. Checks to make Correlate timestamps with cron and systemd timers, process accounting, application logs, and traffic. Identify the process before changing limits. Check whether jobs overlap or scan growing directories, then reschedule, optimize,...

Manual startup can hide missing systemd dependencies, environment variables, mount timing, or permissions. Checks to make Inspect systemctl status and boot journal, verify unit and mounts, check service-user permissions, and define real dependencies. Keep secrets out of unit files. When to get help Linux administration can make startup reproducible...

When a Linux server's clock drifts, authentication tickets may be rejected, certificates may appear invalid, and event timelines can become unreliable. Manually setting the clock may provide temporary relief while leaving the synchronization problem unresolved. Check synchronization status Review the active time service, configured sources, offset, and...

A Linux service can fail during boot when it tries to reach DNS, a mounted filesystem, or a remote database before that dependency is ready. Adding a fixed sleep may hide the race on one host while leaving it on another. Identify the dependency and failure Review systemctl status , the unit file, and journal messages from startup. Determine whether...

A server may report that it cannot create files while df -h still shows free disk space. The filesystem may have run out of inodes, which track file and directory entries. Mail queues, cache directories, sessions, or application-generated temporary files can create many small objects. Check inode usage Use df -ih to identify the full filesystem....

After a server rebuild or administrator change, an SSH key that worked before may be rejected. The cause can be an incorrect username, missing public key, strict home-directory permissions, a changed SSH configuration, or a host key that no longer matches the client record. A host-key warning deserves particular care: confirm the server's identity...

A Linux service may fail after a package update changes a library, configuration default, or service dependency. Downgrading the package can be a temporary recovery step, but it may not reverse a database migration or a config change created by the new version. Identify the change Check package manager history, service logs, dependency versions, and...