Systems Administrator exam: what is tested and how to prepare

Areas the questions come from

• Operating systems: installing and maintaining Windows Server and Linux, services, processes and logs.
• Directory services: Active Directory users, groups, organisational units and Group Policy.
• Linux administration: users and permissions, systemctl and journalctl, package management, cron and SSH.
• Storage: RAID levels, file systems, disk space and permissions on shared folders.
• Backup and recovery: backup types, retention, restore testing and recovery objectives.
• Networking for servers: IP configuration, DNS records, DHCP and the ports common services listen on.
• Security: patching, least privilege, multi-factor authentication and hardening a new server.
• Virtualisation and cloud: hypervisors, virtual machines, snapshots and the basics of cloud services.

What a service needs in order to work

Many troubleshooting questions become easy once you ask what a service depends on. A user logging in to a domain needs the network, a reachable domain controller, working DNS to find that controller, a clock close enough to the domain's time for Kerberos authentication, and an account that is enabled and unlocked. When a question describes a failed login, walk down that list and look for the clue that breaks one link.

The same approach works on Linux. A web service that will not start may have a syntax error in its configuration, a port already in use, missing permissions on its files, or a full disk. The logs usually say which, and that is why reading journalctl output is worth practising.

Backup questions are really about recovery

Questions on backups rarely test whether you know the word "incremental". They test whether you understand what a restore will involve. A full backup copies everything; an incremental copies changes since the last backup of any kind; a differential copies changes since the last full backup. Restoring from incrementals needs the full backup plus every incremental since, while a differential restore needs only the full backup and the latest differential.

Know the 3-2-1 rule: three copies of the data, on two different types of media, with one copy kept off-site. Know the difference between the recovery point objective, how much data the organisation can afford to lose, and the recovery time objective, how long it can afford to be down.

Preparing across two operating systems

Candidates who work mainly with Windows tend to lose marks on Linux, and the reverse. Look honestly at which one you use less, and give it more of your study time. A virtual machine running a common Linux distribution and an evaluation copy of Windows Server give you somewhere to practise both without risk.

Practise the everyday tasks with your own hands: create users and groups, set permissions on a shared folder, schedule a job, restore a file from a backup, and read a log to find why a service failed. Questions written by people who do this work reward people who have done it.

Managing the clock

Knowledge questions on ports, commands and RAID levels should take seconds, so answer them quickly and bank the time. Spend it on scenario questions, where you need to read carefully and eliminate options. If a question about a command option you have never used is taking too long, make your best choice, flag it and move on.