Two RHCSA Mock Exam Papers
Two full 2.5-hour RHCSA mock papers with every ambiguous value pinned, plus a hostile third variant, and a grader that scores tasks fully correct out of 300 against the real 210 pass mark.
Two complete papers, sat under real conditions. Every value is pinned — usernames, UIDs, GIDs, sizes, ports, profile names — so that grading is deterministic and you cannot argue with the result.
Part of RHCSA to RHCE - A 60-Day Lab Curriculum. Graders:
./verify ex200-a --spec and ./verify ex200-b --spec print these papers;
sudo ./verify ex200-a --after-reboot scores them
(the harness).
The conditions
- 2.5 hours, timer visible.
- Closed book — only
man,--helpand/usr/share/doc. No search engine, no notes, no AI. - Revert the node to its
cleansnapshot first. - Graded after a reboot. Anything that does not survive scores zero.
That last point is not a technicality. Red Hat exams are performance based and graded on the state of the machine after a restart, so grade yourself the same way. Reboot, then check.
How to score it
The grader runs one section per task, and a task scores only if every check in it passes. That is how Red Hat grades: partial credit does not exist for a task whose configuration does not survive a reboot.
./verify ex200-b --spec > ~/paper.txt # print the paper
# ... 2.5 hours ...
sudo reboot
sudo ./verify ex200-b --after-reboot # score it
Output is tasks fully correct, scaled to 300 against the real 210 pass mark (70%).
The go/no-go rule. Two consecutive mocks at 85% or above, with everything surviving a reboot, means sit the exam. Below 70% means move the booking by a week. Sitting it unprepared costs more time than rescheduling.
Mock A
- Set the hostname to
server-a.lab.local, permanently. - Configure a static IP on the primary interface with two DNS servers and the
search domain
lab.local. It must survive a reboot. - Group
finance, GID 6000. Usersfin1andfin2in it.fin2must not be able to log in interactively. fin1's password expires in 45 days, with a 10-day warning./srv/finance— group-owned byfinance, group-writable, users cannot delete each other's files, new files inherit the group.- Read-only ACL for user
auditoron/srv/finance, inherited by new files, with no change to group membership. - A 3 GB XFS filesystem from the blank disk, mounted at
/financeby UUID, withnodev. - 1 GB of swap, persistent.
- Volume group
vgappwith 32 MB extents; a 2 GB logical volume mounted at/appdata; then extend it to 5 GB with the filesystem grown. - httpd installed and enabled, serving from
/finance/www, SELinux enforcing. - http open in the firewall, permanently.
- A systemd timer running
/usr/local/bin/audit.shdaily at 03:00, persistent. - The journal persistent and capped at 300 MB.
fin2denied the use of cron.- Reset the root password from GRUB (simulate the lockout).
Mock B
- A repository named
locallabfrom the local directory/repo, gpgcheck off, and a package installed from it. - Three users:
ops1with UID 3101,ops2, andsvc1with UID 3199 and no interactive shell. /srv/shared— every new file created there is group-owned bydevelopersautomatically.- A default ACL on
/srv/sharedgiving userqa1write access to new files. - Volume group
vgdataon a blank disk; logical volumelvapp, XFS, mounted at/app; extended by 2 GB to 6 GB with the filesystem grown online. - Logical volume
lvlogs, ext4, mounted at/applogs, reduced to 1 GB safely. - The second node exports
/srv/share; this node mounts it persistently at/mnt/share. - autofs mounts that share at
/net/dataon demand, with a 60-second timeout. - A rootless container serving a page on port 8080 from
/opt/webdata, auto-started at boot via a systemd--userunit with lingering enabled. - sshd moved to port 2222, with SELinux enforcing and the firewall open.
- Every SUID file on the system listed in
/root/reports/suid.txt. /usr/local/bin/unitcheck— exits non-zero when any systemd unit has failed.- The tuned profile
virtual-guest, set persistently. - A non-default kernel set as the permanent default.
/etc/fstabarranged so a failed mount does not prevent the machine booting.
Mock C — hostile
Mock B, on a machine you have already wrecked. Repair the estate first, then complete the tasks, in the same 2.5 hours.
virsh snapshot-create-as rhel01 pre-mock
sudo ./break.sh all # ten single-cause faults
sudo ./break/break-advanced.sh combo --yes # three harder ones
sudo reboot
Grade with all three: ./verify breakfix, ./verify advanced, ./verify ex200-b.
If Mock C is comfortable, you are ready. Inheriting a broken estate and being asked to configure it further is much closer to both the exam and the job than a clean machine ever is.
The faults are described in RHCSA Security Labs - SELinux firewalld and Boot Recovery and Ten Advanced RHEL Break-Fix Drills.
Recording results
| Mock | Date | Score | Time used | Weak areas |
|---|---|---|---|---|
| A | ||||
| B | ||||
| C |
The "weak areas" column is the only part of this table that matters. Two days before the exam, work only on what these mocks proved is weak — and do not start a new topic, because cramming something unfamiliar in the final week displaces something you already know.
What the papers deliberately include
Each paper contains at least one of every mistake that turns a working configuration into a zero:
- A service that must be enabled, not merely started (A10, B13).
- A mount that must be by UUID (A7), and one that must not block boot (B15).
- A SELinux context that must survive a relabel, not just a reboot (A10, B10).
- A firewall rule that must be permanent (A11, B10).
- A timer that must be
Persistent=true(A12). - A
--userunit that needs lingering to start at boot (B9). - An ACL that must be inherited by new files (A6, B4).
If you can name which requirement in each paper hides which trap, you have understood the exam better than most of the material available for it.
→ Full protocol: Red Hat Exam Day Protocol