System Administration
Core duties
Provisioning and de-provisioning accounts, patching, backups, monitoring, and documentation. Least privilege and role-based groups matter more than individual permissions. Change management, meaning recording what you changed, why, and how to roll back, prevents most outages from becoming long ones.
Backups
Follow 3-2-1: three copies, two media types, one offsite. RAID protects against disk failure, not deletion, ransomware, or fire. Snapshots are not backups unless stored separately. Test restores; an untested backup is only a hope.
Monitoring and logs
Watch disk space, CPU, memory, uptime, and service health. Logs from Event Viewer, journalctl, or syslog are evidence; check timestamps and correlate events. Scheduled tasks and cron automate maintenance.
Traps
- Treating RAID or snapshots as a backup.
- Granting admin rights to 'fix' a permission problem.
- Disabling a firewall or antivirus to make something work.
- Making several changes at once, then not knowing which one fixed it.
- Forgetting to remove access for departed staff.
- Skipping documentation.
Rule of thumb
Before any risky change: know the rollback, announce the maintenance window, back up the configuration, change one thing, verify, and record it. If a server fills its disk, safely clear logs or temp files, then find what is filling it rather than only adding space.