Question 570
A storage controller is visible in lspci but no /dev block node appears. Which subsystem is the next evidence source?
- Inspect BMC power and hardware sensor events
- Cooling, airflow and fan health
- CPU, memory, or platform hardware fault
- Nested virtualization exposure
- Controller driver binding and dmesg errors
Explanation
Correct Answer
E:PCI enumeration proves the hardware is discovered, not that a driver has initialized attached storage. Kernel messages reveal binding and device-probe failures.
Incorrect Answers
A:Out-of-band management can report chassis power, thermal and hardware faults even when the OS is unavailable; the evidence calls for explaining a PCI controller with no disk nodes, not merely this separate operation.
B:Thermal throttling reduces frequency to protect silicon; the task involves explaining a PCI controller with no disk nodes, whereas this addresses another subsystem.
C:Machine-check records can report low-level hardware errors rather than application exceptions; its scope differs from explaining a PCI controller with no disk nodes, so it cannot meet the stated requirement.
D:Guests need virtualization extensions exposed by the host hypervisor to run nested accelerated VMs; that is not equivalent to explaining a PCI controller with no disk nodes.
Question 571
A process remains visible in state Z but consumes no CPU after a child exit. What should be repaired?
- Uninterruptible sleep waiting on kernel I/O
- Invalid process memory access
- Parent process child-reaping behavior
- Installed shared libraries and dynamic loader paths
Explanation
Correct Answer
C:A zombie persists until its parent collects exit status or it is reparented and reaped. Killing the already exited child is ineffective.
Incorrect Answers
A:Processes in uninterruptible sleep cannot complete signal handling until the kernel wait ends; it answers a neighboring question rather than clearing unreaped child exit records.
B:A segmentation fault occurs when a process accesses memory in violation of protection rules; that changes a different part of the system; the requested outcome is clearing unreaped child exit records.
D:The executable cannot start when a required shared object cannot be located or loaded; it will not provide clearing unreaped child exit records under these conditions.
Question 572
A service cannot write to /srv even though the owner permission bits are correct, and quota reports limit reached. What is the likely blocker?
- Container writable-layer or quota capacity
- NFS server or network path availability
- Processes holding files or working directories on the mount
- Grow the filesystem on the extended LV
- Kernel remounted filesystem read-only after errors
- Filesystem user or project quota enforcement
Explanation
Correct Answer
F:Quotas may deny additional allocation even when the filesystem has free blocks and ordinary permissions allow writes. Review the quota scope and usage.
Incorrect Answers
A:A container's storage driver or configured filesystem quota can limit writes independently of other host mounts; the requirement is diagnosing quota-enforced write denial, outside the effect of this choice.
B:A hard network mount can wait on unavailable remote I/O; it does not deliver diagnosing quota-enforced write denial, which is decisive in this scenario.
C:Open file descriptors and current working directories can keep a filesystem busy; the evidence calls for diagnosing quota-enforced write denial, not merely this separate operation.
D:Increasing the LVM block device does not always resize the filesystem; the task involves diagnosing quota-enforced write denial, whereas this addresses another subsystem.
E:Some filesystems remount read-only after serious consistency or storage faults; its scope differs from diagnosing quota-enforced write denial, so it cannot meet the stated requirement.
Question 573
Linux cannot unmount /backup because the target is busy. What should be identified before forcing anything?
- Available filesystem inodes
- Failing storage hardware
- Create a supported filesystem before mounting
- Processes holding files or working directories on the mount
Explanation
Correct Answer
D:Open file descriptors and current working directories can keep a filesystem busy. Locate holders with fuser or lsof and stop them safely.
Incorrect Answers
A:A filesystem may exhaust its inode inventory while still showing free data blocks; inode counts require a separate check. df -i distinguishes inode exhaustion from block capacity; that is not equivalent to finding why an unmount reports busy.
B:Increasing device errors and sector remapping suggest media degradation; the administrator instead needs finding why an unmount reports busy; this option serves a different purpose.
C:A block device with no recognized filesystem cannot be mounted as ext4 or XFS until formatted appropriately; it answers a neighboring question rather than finding why an unmount reports busy.
Question 574
A UEFI system cannot find its Linux bootloader after a disk migration, but the EFI partition still contains the files. Which configuration is suspect?
- Initramfs with the needed storage module
- /etc/fstab entry for /data
- UEFI firmware boot entry or EFI path
- systemctl list-dependencies and failed units
Explanation
Correct Answer
C:UEFI firmware entries locate an EFI executable by device and path. Migration can leave stale entries even when files still exist.
Incorrect Answers
A:A driver only available on the inaccessible root filesystem arrives too late; it will not provide fixing firmware references to an existing EFI loader under these conditions.
B:Incorrect identifiers, mount types or required mounts can push boot into emergency mode; using it would leave fixing firmware references to an existing EFI loader unresolved.
D:Systemd dependency inspection reveals which required unit blocked target activation; it does not deliver fixing firmware references to an existing EFI loader, which is decisive in this scenario.
Question 575
A service receives SIGTERM yet continues for several minutes and misses its stop timeout. What should be reviewed?
- Service shutdown handler and TimeoutStopSec
- Installed shared libraries and dynamic loader paths
- Invalid ExecStart path or inability to execute it
- Uninterruptible sleep waiting on kernel I/O
Explanation
Correct Answer
A:Slow cleanup or a stuck process can exceed systemd's grace period. Review application logs and service stop timeout rather than forcing every stop immediately.
Incorrect Answers
B:The executable cannot start when a required shared object cannot be located or loaded; its scope differs from resolving graceful stop timeout failures, so it cannot meet the stated requirement.
C:Systemd's 203/EXEC commonly indicates the configured executable could not be started; that is not equivalent to resolving graceful stop timeout failures.
D:Processes in uninterruptible sleep cannot complete signal handling until the kernel wait ends; the administrator instead needs resolving graceful stop timeout failures; this option serves a different purpose.
Question 576
A mounted ext4 filesystem has checksum-like metadata errors after a sudden power loss. What recovery practice is most appropriate?
- Filesystem user or project quota enforcement
- Verify md array state and replace the failed member
- Failing storage hardware
- Offline filesystem consistency check after backup
Explanation
Correct Answer
D:Repair tools may alter filesystem structures, so preserve data and unmount the affected filesystem first. Avoid blindly running write-mode repair against a live mount.
Incorrect Answers
A:Quotas may deny additional allocation even when the filesystem has free blocks and ordinary permissions allow writes; that changes a different part of the system; the requested outcome is repairing damaged filesystem metadata safely.
B:RAID1 can remain active in degraded mode while mirror redundancy is lost; this does not achieve repairing damaged filesystem metadata safely on the system described.
C:Increasing device errors and sector remapping suggest media degradation; it will not provide repairing damaged filesystem metadata safely under these conditions.
Question 577
A program crashes with segmentation fault when handling one malformed input. Which category of fault fits?
- Parent process child-reaping behavior
- Invalid ExecStart path or inability to execute it
- User= identity in the service unit
- Uninterruptible sleep waiting on kernel I/O
- Invalid process memory access
Explanation
Correct Answer
E:A segmentation fault occurs when a process accesses memory in violation of protection rules. Input-specific reproducibility can assist debugging.
Incorrect Answers
A:A zombie persists until its parent collects exit status or it is reparented and reaped; it does not deliver interpreting a segmentation fault from a program, which is decisive in this scenario.
B:Systemd's 203/EXEC commonly indicates the configured executable could not be started; the evidence calls for interpreting a segmentation fault from a program, not merely this separate operation.
C:Systemd switches to the configured identity before starting a process; the task involves interpreting a segmentation fault from a program, whereas this addresses another subsystem.
D:Processes in uninterruptible sleep cannot complete signal handling until the kernel wait ends; its scope differs from interpreting a segmentation fault from a program, so it cannot meet the stated requirement.
Question 578
PID 1 fails to reach multi-user.target because one service has a failed requirement. Which systemd inspection should be prioritized?
- Last kernel console messages and panic context
- Different PATH and execution environment
- systemctl list-dependencies and failed units
- Actual filesystem UUIDs against the root boot argument
Explanation
Correct Answer
C:Systemd dependency inspection reveals which required unit blocked target activation. A generic reboot would not identify the dependency failure.
Incorrect Answers
A:Panic logs, call traces and boot parameters narrow the failing driver or subsystem; the administrator instead needs finding the unit dependency that blocked boot; this option serves a different purpose.
B:Systemd services receive a controlled environment that may omit interactive profile additions; it answers a neighboring question rather than finding the unit dependency that blocked boot.
D:A stale UUID can point to a nonexistent root volume; this does not achieve finding the unit dependency that blocked boot on the system described.
Question 579
After a package update, a shared-library application fails with loader cannot open shared object file. Which check is most relevant?
- Service shutdown handler and TimeoutStopSec
- Installed shared libraries and dynamic loader paths
- Uninterruptible sleep waiting on kernel I/O
- Invalid ExecStart path or inability to execute it
Explanation
Correct Answer
B:The executable cannot start when a required shared object cannot be located or loaded. Package dependency repair or library-path correction may be needed.
Incorrect Answers
A:Slow cleanup or a stuck process can exceed systemd's grace period; using it would leave diagnosing runtime shared-library resolution unresolved.
C:Processes in uninterruptible sleep cannot complete signal handling until the kernel wait ends; it does not deliver diagnosing runtime shared-library resolution, which is decisive in this scenario.
D:Systemd's 203/EXEC commonly indicates the configured executable could not be started; the evidence calls for diagnosing runtime shared-library resolution, not merely this separate operation.
Question 580
A RAID1 array shows a failed device but the volume remains accessible. What should the operator prioritize?
- Verify md array state and replace the failed member
- Available filesystem inodes
- Create a supported filesystem before mounting
- Failing storage hardware
Explanation
Correct Answer
A:RAID1 can remain active in degraded mode while mirror redundancy is lost. Array health and a controlled rebuild should be confirmed promptly.
Incorrect Answers
B:An inode shortage prevents creating new files despite free block capacity, so inode utilization must be checked independently. df -i distinguishes inode exhaustion from block capacity; that is not equivalent to restoring redundancy in a degraded mirror.
C:A block device with no recognized filesystem cannot be mounted as ext4 or XFS until formatted appropriately; the administrator instead needs restoring redundancy in a degraded mirror; this option serves a different purpose.
D:Increasing device errors and sector remapping suggest media degradation; it answers a neighboring question rather than restoring redundancy in a degraded mirror.