Skip to content

Per-exam playbooks

LFCS study plan: the five domains, a distribution-agnostic exam and loop-device practice

LFCS, the Linux Foundation Certified System Administrator, is 17-20 performance-based tasks in 120 minutes with a 67% pass mark. It is the one Golden Kubestronaut exam with no Kubernetes in it, and the only one where the web is closed: man pages and the distribution's own documents are all you get. This guide covers the five domains, what "independent of distribution-specific tasks" means for your practice, how to rehearse storage work safely on loop devices, and a six-week plan.

By Ibrahim YesilayUpdated 7 min read

Key takeaways

  • Operations Deployment and Networking are 25% each; together they are half the exam.
  • No web documentation is allowed: only man pages, documents under /usr/share, and packages from the distribution.
  • The exam is independent of distribution-specific tasks, so practise the tools that behave the same everywhere: systemd, ip, LVM, ssh, nftables.
  • Loop devices let you practise partitioning, LVM, swap and filesystem repair on files instead of real disks.
  • Read the man page of a tool once before you need it on exam day; finding the right section quickly is a skill.
On this page

LFCS in numbers

The exam is performance-based: you solve tasks on the command line of Linux hosts reached over ssh. The candidate instructions give 17-20 performance-based tasks and a 67% pass mark; the certification page gives 120 minutes, a $445 list price and 2-year validity. The current outline has been in effect since 11 May 2023.

DomainWeightCompetencies
Operations Deployment25%Kernel parameters, processes and services, scheduled jobs, packages and repositories, recovery from failures, libvirt virtual machines, containers, SELinux
Networking25%IPv4 and IPv6, hostname resolution, time synchronisation, troubleshooting, OpenSSH, packet filtering, port redirection and NAT, static routes, bridges and bonds, reverse proxies and load balancers
Storage20%LVM, the virtual file system, filesystems, remote filesystems and network block devices, swap, automounters, storage performance
Essential Commands20%Git basics, services, performance monitoring, application constraints, disk space, SSL certificates
Users and Groups10%Local accounts, environment profiles, resource limits, ACLs, LDAP accounts

What a distribution-agnostic exam means

Since the May 2023 update, the exam content is independent of distribution-specific tasks, and you no longer choose a platform when you prepare. In practice the tasks describe an outcome, such as "make this filesystem mount at boot" or "allow only these ports", and it is up to you to reach it with the tools on the host. That favours the tools that are the same on every mainstream distribution, and it means you should know how to find out what the host gives you. In practice, learn the tools that behave the same everywhere first, such as systemctl, journalctl, ip, ss and the core file utilities, and treat package-manager differences as a detail to check with man pages.

  • Learn the cross-distribution tools first: systemctl and journalctl, ip, ss, lvm commands, mount and fstab, ssh and sshd_config, cron and systemd timers.
  • Know both package families at the level of install, query and search: apt and dpkg, dnf and rpm.
  • Find out what the host uses before changing it: which firewall front end, which network configuration tool, which MAC system.
  • The instructions warn that installing services may require changing security policies; check SELinux or AppArmor and the firewall when a service will not start or answer.

Working without the web

LFCS is the strictest exam on resources. The allowed list is man pages, the documents a distribution installs under /usr/share, and packages from the distribution, which you may install yourself. There is no search engine and no project website. The SSH hosts have Emacs, Git, Nano, sudo and vim; the base system you start on has none of them.

  1. Practise answering questions with man -k keyword and man section numbers, for example man 5 fstab against man 8 mount.
  2. Look under /usr/share/doc for the example configurations packages ship.
  3. Use the --help output of a command before opening its man page; it is faster for flags you half remember.
  4. Rehearse a whole task with the browser closed at least once a week.

Practising storage on loop devices

Storage is 20% of the exam, and you do not want to practise repartitioning on the disk your system boots from. Loop devices solve that: a file on disk is attached as a block device, and every storage command treats it like a real disk. You can partition it, add it to a volume group, format it, use it as swap, corrupt it and repair it, then detach it and delete the file.

  1. Create two empty files, for example with truncate -s 1G disk1.img and disk2.img.
  2. Attach them with sudo losetup -fP --show disk1.img and note the /dev/loopN name it prints.
  3. Build on them: pvcreate, vgcreate, lvcreate, mkfs, then a mount and an fstab entry by UUID.
  4. Extend a logical volume and grow the filesystem online; then shrink a test filesystem to see which types allow it.
  5. Detach with losetup -d when finished and remove the fstab entry so the next boot does not wait for it.

Operations and networking: half the exam

Operations Deployment and Networking are 25% each. Operations is about keeping a system running: kernel parameters set now and at boot, services that start and survive a reboot, scheduled jobs, packages and repositories, containers and libvirt virtual machines, and recovering from failures. Networking covers addressing and name resolution, time, OpenSSH, packet filtering and NAT, static routes, bridges and bonds, and reverse proxies and load balancers.

  • Set a sysctl value persistently in /etc/sysctl.d and apply it without rebooting.
  • Fix a failing systemd unit, enable it, and confirm it starts after a reboot.
  • Schedule a job with a systemd timer and with cron, and read when it will next run.
  • Configure sshd for key-only login and a different port, then reload it without locking yourself out.
  • Write packet filter rules that allow a short list of ports, and a port redirect.
  • Put a reverse proxy in front of two local backends and check that requests alternate.

Essential commands, users and groups

Essential Commands (20%) mixes Git basics, services, performance and disk-space troubleshooting, application constraints and SSL certificates. Users and Groups (10%) is the smallest domain but the quickest points: local accounts with specific IDs and expiry, environment profiles, resource limits, ACLs and LDAP-backed accounts.

  • Clone, branch, commit and push to a local bare repository.
  • Find what is filling a disk and free space safely.
  • Create a self-signed certificate and key and read its subject and dates with openssl.
  • Set default ACLs and the setgid bit on a shared directory.
  • Configure a per-group resource limit and verify it in a new session.

A six-week LFCS plan

The plan follows the weights and keeps one closed-book session each week. Use a disposable virtual machine you can rebuild quickly, and attach loop devices for every storage exercise. Take the included simulator in week six.

LFCS hands-on labs are in preview: the GoldenKube CLI runs its scenarios on your own machine, but it has no public release yet.

WeekFocus
1Essential commands, man pages, services and journald; users, groups and ACLs
2Operations: sysctl, systemd units and timers, cron, packages and repositories
3Storage on loop devices: partitions, LVM, filesystems, swap, fstab, automount
4Networking: addressing, resolution, time sync, OpenSSH, static routes
5Packet filtering and NAT, bridges and bonds, reverse proxy; containers, libvirt, SELinux
6Recovery drills, closed-book timed tasks, the simulator; book the exam

Habits for a Linux performance exam

LFCS tasks are checked on the system after you finish, so a change that works now but disappears later, or that lives in the wrong place, can score nothing. The habits below apply to almost every domain and are worth practising until they are automatic.

  1. Make changes persistent: enable services as well as starting them, and write mounts to /etc/fstab as well as mounting them.
  2. Test a persistent change the way the system will use it, for example with mount -a after editing /etc/fstab.
  3. Use the exact names, paths, users and permissions in the task text; a correct result under another name is a different result.
  4. Find unknown commands with man -k or apropos, then read the EXAMPLES section of the man page first.
  5. Look under /usr/share/doc for configuration examples a package installs.
  6. Check the result as an ordinary user when the task is about permissions or access.

Frequently asked questions

Which Linux distribution does the LFCS exam use?

The Linux Foundation does not name one. The certification page says the exam is independent of distribution-specific tasks, and the May 2023 update removed the need to choose a platform in the exam preparation checklist. Practise on one mainstream distribution, and know the equivalents on the other family where commands differ.

Can I use the internet during LFCS?

No. The resources-allowed page lists man pages, documents installed by the distribution under /usr/share, and packages that are part of the distribution, which you may install. Resources that cannot be reached from the terminal are not allowed.

Is LFCS required for Golden Kubestronaut?

Yes. The CNCF FAQ says Golden Kubestronaut requires all CNCF certifications and LFCS. LFCS is not required for the five-exam Kubestronaut title.

Should I take LFCS before CKA?

It is not required; LFCS has no prerequisite and CKA requires none either. Taking it first is a reasonable choice if you are not yet fast at the shell, because every Kubernetes hands-on exam assumes those skills.

Does LFCS include a simulator?

The certification page lists an exam simulator among the items included with the exam purchase. Use it late in your preparation, once the domains are covered.

How is SELinux practised on a distribution that uses AppArmor?

Only partially. The curriculum lists creating and enforcing mandatory access control with SELinux. If your practice machine uses AppArmor, learn the SELinux concepts and commands (modes, contexts, file context rules) on a RHEL-family virtual machine as well, because they do not transfer one to one.

Sources

Every exam fact on this page comes from these pages, read on the date shown.

Independent community project, not affiliated with or endorsed by CNCF or The Linux Foundation.