Linux Admin Scenario Questions: Seven Worked Incidents
For a Linux incident scenario, explain the first evidence you need, the decision it supports and how you would verify an authorised correction. These seven original cases cover boot failure, storage exhaustion, SSH access, high load, a failed service, deleted data and lost administrative access. Avoid treating every symptom as permission to restart, repair a disk or disable a security control.
Topic: Linux Administration in IT Infrastructure · Interview preparation
Quick answer: For a Linux incident scenario, explain the first evidence you need, the decision it supports and how you would verify an authorised correction. These seven original cases cover boot failure, storage exhaustion, SSH access, high load, a failed service, deleted data and lost administrative access. Avoid treating every symptom as permission to restart, repair a disk or disable a security control. State the environment and business impact, preserve useful evidence and use the supported recovery path for the actual distribution, filesystem and service.
Key takeaways
- Identify the environment and last known working state.
- Choose the next check from the symptom and available evidence.
- Preserve logs and required data before changing state.
- Use filesystem-specific recovery with the correct prerequisites.
- Define a successful application result after any correction.
Table of Contents
- Case 1: What do you do when a server will not boot?
- Case 2: What if an application reports no space left?
- Case 3: Why does an SSH connection fail?
- Case 4: Does high load prove the CPU is saturated?
- Case 5: How do you approach a failed systemd service?
- Case 6: What happens after important data is deleted?
- Case 7: What if nobody can use the administrative account?
- References
Case 1: What do you do when a server will not boot?
Question: after approved maintenance, a fictional server stops before its normal login screen. Where do you begin? Answer cue: determine how far startup reached and obtain console evidence. Firmware detection, boot loader, kernel, initramfs and service startup are different stages. Record the exact message and recent change before choosing a recovery procedure.
The Red Hat rescue-environment guidance is a distribution-specific reference, not permission to run a generic repair command against any disk. Identify the device, filesystem, mount state and data protection prerequisites. If the message points to a missing configured mount, investigate that dependency rather than declaring the disk dead. Restoring login is one check; also verify the services and data the server is supposed to provide.
Case 2: What if an application reports no space left?
Question: the application's data write fails, but a colleague says there is free space somewhere on the server. Which question comes first? Answer cue: identify the filesystem backing the failed path. The GNU df manual describes filesystem space and inode reporting; free capacity on a different mount does not resolve this write.
The GNU du manual describes usage for selected paths. Compare scope and units before reconciling the views. Check whether blocks, inodes, quota or another relevant limit explains the error. In the invented case, a report spool has many small files on its own mount. Do not delete logs or user data blindly. The authorised owner must choose retention, capacity or application corrections, then verify a representative write and the intended service behaviour.
Case 3: Why does an SSH connection fail?
Question: a user says “SSH is down.” What must you clarify? Answer cue: distinguish name-resolution failure, timeout, refused connection, unexpected host identity and rejected user authentication. These are different observations. Establish the intended host, port and approved access route, then inspect the relevant client and server evidence.
The OpenSSH manuals separate connection and authentication functions. A successful ping cannot establish that the required SSH listener is reachable or that the account is authorised. In the fictional case, console access shows the service is available only on the intended management interface. Compare the user's route with that design. Do not bypass host-key warnings, open broad network access or restart the service without a supported reason and an authorised change plan.
Case 4: Does high load prove the CPU is saturated?
Question: load rises while the displayed CPU use is modest. Is the measurement inconsistent? Answer cue: Linux load average includes runnable work and tasks in uninterruptible waiting, as the load-average manual explains. Read CPU, memory, I/O and application observations over a relevant interval rather than treating load as a CPU percentage.
In the invented case, requests slow after a storage-path change. A useful next step examines the timing of application waits and storage evidence; the supplied facts do not prove the device itself failed. A single snapshot can miss a short burst. State which observation would distinguish CPU demand from waiting or another constraint. Avoid recommending extra CPU or changing kernel tuning simply because the load number looks large.
| Observation | Useful next distinction |
|---|---|
| High load, low CPU | Runnable work versus uninterruptible waits |
| Slow requests | Application stage and matching time interval |
| Memory pressure concern | Active pressure versus a historical allocation |
| Recent storage change | Correlation versus established cause |
Case 5: How do you approach a failed systemd service?
Question: a service fails immediately after a configuration update. What would your first answer include? Answer cue: the unit's current state, relevant error, precise change and application-specific validation route. The systemctl reference documents unit state, while the journalctl reference explains filtering the journal by unit and time.
In the fictional record, the service reports that its configured data directory cannot be accessed. Check the effective service identity and the intended path before assuming the package must be reinstalled. The pathname-resolution manual explains directory search permissions, and the RHEL SELinux guide covers an additional access-control layer in that environment. Confirm a denial's cause before correcting labels or policy. Disabling enforcement is not the default answer to an unexplained failure.
Case 6: What happens after important data is deleted?
Question: a user reports deleting a required file. What do you ask before attempting recovery? Answer cue: establish the path, deletion time, filesystem or service, importance and available backup or version history. Preserve the situation through the incident or recovery process, including limiting further affected storage activity when the responsible specialist requires it.
Do not promise that a universal undelete command can recover the material. Continued writes, storage behaviour and the application's own state can affect the available options. In the invented case, a required spreadsheet exists in an approved versioned backup. The owner restores it to a controlled alternate location and checks contents and business suitability before replacement. That would validate this recovery route; it would not establish that every deleted file on the server is recoverable.
Case 7: What if nobody can use the administrative account?
Question: a team has lost administrative access to its server. Does knowing a password-reset technique settle the problem? Answer cue: first establish ownership, approved console or recovery access, identity controls and the actual failure. A locked account, lost secret, failed identity service and broken network path need different responses.
Use the supported distribution and platform recovery process with the system owner's authority. Protect credentials and preserve an audit record. For the fictional team, the chosen recovery route depends on an approved console and a known backup administrator; the exercise supplies no bypass or reset sequence. After access is restored, verify required controls and application health, then fix the documented access-recovery gap so the next authorised administrator has a reliable route.
| Case review | Evidence to retain |
|---|---|
| Before action | Affected work, observation and last working state |
| Decision | Hypothesis and authorised supported procedure |
| After action | Access or service result and remaining fault |
| Follow-up | Owner of the underlying maintenance or recovery gap |
Related reading: Linux diagnostic question bank; Beginner Linux concepts.
References
- Red Hat filesystem repair in rescue environments
- GNU df manual hosted by Linux man-pages
- GNU du manual hosted by Linux man-pages
- OpenSSH manual pages
- Linux load average manual
- Upstream systemctl reference
- Upstream journalctl reference
- Linux pathname resolution manual
- Red Hat Enterprise Linux 10 SELinux guide
Frequently Asked Questions
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Wow
0
Sad
0
Angry
0