System Management practice questions

chmod a+x script.sh

chmod a+x adds execute permission for user, group and other, changing the mode from 644 to 755. This makes the script executable by all users, satisfying the stem's requirement, whereas chmod +x alone would be subject to umask.

0022

Setting umask 0022 strips group and other write bits from the default 666 file and 777 directory modes, yielding 644 files and 755 directories. This satisfies the stem's exact permission requirement, since umask bits are subtracted from the base modes the kernel assigns at creation.

du -h /

du -h / walks the root directory tree and reports disk usage per directory in human-readable units, letting an administrator identify space consumption. It complements df by showing where space is used, though it reports paths rather than mounted filesystems.

df -h

df reports filesystem-level usage for every mounted filesystem, and the -h flag renders sizes in human-readable units. This directly satisfies the stem's requirement to display disk space consumed per mounted filesystem, unlike du, which reports directory consumption.

rescue.target

rescue.target starts a single-user shell with local filesystems mounted but networking disabled, satisfying the minimal-processes, no-network constraint. It is the gentler of the two minimal targets, allowing basic repair work without bringing up network services.

emergency.target

emergency.target provides the most minimal environment: a single-user shell with only the root filesystem mounted read-only and no network services started. It satisfies the requirement for a minimal process set with networking disabled, being more restrictive than rescue.target.

systemctl start service && systemctl enable service

Chaining start then enable launches the unit in the current session and creates the boot-time symlink, achieving both goals in one line. The order matters only for immediate availability; enable alone would defer activation until the next reboot.

systemctl enable --now service

Correct: systemctl enable --now service is a valid systemctl command that both enables the service for boot and starts it immediately.

find . -type f -size +1M

Correct. This command correctly lists regular files larger than 1 MB. The -type f ensures only regular files are considered, and -size +1M matches files larger than 1 MB.

find . -size +1M -type f

Correct. Although the order of -size and -type is reversed, the command is still syntactically correct and will produce the same result as option A. The find command evaluates predicates left-to-right, but both forms work because all conditions must be met regardless of order.

find /var/log -mtime -7 -size +10M

-mtime -7 matches files modified within the last seven days, while -size +10M matches files larger than ten megabytes. Combining both predicates with implicit AND on /var/log returns exactly the files meeting both constraints in a single pass.

find /var/log -mmin -10080 -size +10M

Correct. This is functionally equivalent to option A; -mmin -10080 matches files modified in the last 10080 minutes (7 days), and -size +10M matches files larger than 10 MB.

grep -in 'error' logfile

`grep -in 'error' logfile` satisfies both constraints: `-i` matches 'ERROR' regardless of case, and `-n` prefixes each matching line with its line number. Combined, they return the first occurrence's line number directly, unlike `-c` (count only) or plain `grep` (no numbering).

chmod u+s file

chmod u+s file correctly adds only the SUID bit for the owner without changing any other permissions, preserving the existing rwxr-x---.

sed -i 's/192.168.1.1/10.0.0.1/g' config.txt

The -i flag performs an in-place edit, and the g suffix replaces every occurrence on each line rather than only the first. Combined with the s/// substitution of the old address for the new one, this satisfies the requirement to replace all occurrences in config.txt.

ps aux

The aux BSD-style option set shows every process regardless of controlling terminal, and the u format adds the full command line with user context. This satisfies the requirement to expose daemonised or malicious processes lacking a tty.

ps -ef

`ps -ef` lists every process on the system, including daemons and others detached from a controlling terminal, satisfying the "without a controlling terminal" constraint. The `-f` flag expands output to show the full command line, including UID, PID, PPID and start time, meeting the requirement to inspect each process's complete invocation.

/bin

The FHS mandates /bin as a required directory because it holds essential user command binaries needed for single-user and boot-time operation. Its presence is compulsory on every conforming Linux system, satisfying the question's requirement for directories that must exist.

/dev

/dev is a required FHS directory because it contains device files representing hardware and pseudo-devices. The kernel and userland depend on it for device access, so its presence is mandatory on any conforming Linux system, matching the question's requirement.

/etc

/etc is required by the FHS because it stores host-specific system configuration files. Applications and services read their settings from here, making it essential for a functioning system and satisfying the question's requirement for directories that must be present.

systemctl status service && journalctl -n 50

systemctl status reports the unit's load, active and failed state, while journalctl -n 50 retrieves the fifty most recent log entries for that service. Together they satisfy the requirement to check status and view recent boot-failure logs under systemd.

rpm -e package

rpm -e removes the named package from the RPM database and deletes its installed files. Configuration files marked as config are preserved by default, so the administrator must verify or add --noscripts and config handling to purge them fully.

Use renice with root privileges via sudo

Lowering a nice value below its current level, especially to a negative value, requires CAP_SYS_NICE. A non-owner cannot do this, so the administrator must invoke renice through sudo to gain root privileges, satisfying the privilege constraint in the stem.

SGID; new files in the directory inherit the group

The leading 2 in 2755 sets the SGID bit. On a directory, SGID forces newly created files and subdirectories to inherit the directory's group ownership rather than the creator's primary group, which is exactly the inheritance behaviour described.

grub2-mkconfig -o /boot/grub2/grub.cfg

grub2-mkconfig reads /etc/default/grub and the scripts in /etc/grub.d/, then writes the regenerated configuration to the path given by -o. On BIOS systems that path is /boot/grub2/grub.cfg, making this the correct regeneration command after editing those files.

awk -F: '{print $2}' passwd

awk's -F: flag sets the input field separator to colon, so '{print $2}' extracts exactly the second field from each line of passwd. This satisfies the requirement to display only the second colon-delimited field, unlike cut alternatives that would need different syntax.

journalctl -u servicename

`journalctl -u servicename` filters the systemd journal by unit, surfacing that service's start attempts, exit codes and dependency failures. It satisfies the stem's requirement to investigate logs for a unit defined in /usr/lib/systemd/system, reading the persistent journal rather than the unit file itself.

systemctl status servicename

`systemctl status servicename` queries systemd for the unit's current state, showing whether it is active, failed, or inactive, plus the most recent log lines and the exit code. This directly satisfies the scenario's need to investigate why a service in /usr/lib/systemd/system fails to start.

journalctl -f

`journalctl -f` streams live journal entries, letting the technician watch messages as the failing unit attempts to start. It satisfies the investigation requirement by revealing real-time errors, though it lacks unit filtering, so output includes unrelated services unless combined with `-u`.

rescue.target

rescue.target starts a minimal single-user environment with essential filesystems mounted and basic services only, ideal for troubleshooting a failing boot. emergency.target is more restrictive, mounting root read-only without normal service startup, while multi-user.target loads the full service set.