Question 600
An NFS client hangs when opening /mnt/share and a remote server is unreachable. Which dependency is probably failing?
- Offline filesystem consistency check after backup
- NFS server or network path availability
- Container writable-layer or quota capacity
- Create a supported filesystem before mounting
Explanation
Correct Answer
B:A hard network mount can wait on unavailable remote I/O. Confirm server reachability and mount options before assuming local disk corruption.
Incorrect Answers
A:Repair tools may alter filesystem structures, so preserve data and unmount the affected filesystem first; that changes a different part of the system; the requested outcome is recognizing remote filesystem access stalls.
C:A container's storage driver or configured filesystem quota can limit writes independently of other host mounts; it will not provide recognizing remote filesystem access stalls under these conditions.
D:A block device with no recognized filesystem cannot be mounted as ext4 or XFS until formatted appropriately; using it would leave recognizing remote filesystem access stalls unresolved.
Question 601
After a kernel update, the newest image panics while the previous kernel still boots from GRUB. What is the safest immediate recovery path?
- Actual filesystem UUIDs against the root boot argument
- Boot the prior known-good kernel
- systemctl list-dependencies and failed units
- UEFI firmware boot entry or EFI path
Explanation
Correct Answer
B:A working older kernel isolates the regression and restores service without immediately rebuilding disks. The failed image can then be investigated offline.
Incorrect Answers
A:A stale UUID can point to a nonexistent root volume; it does not deliver recovering from a newly broken kernel image, which is decisive in this scenario.
C:Systemd dependency inspection reveals which required unit blocked target activation; the task involves recovering from a newly broken kernel image, whereas this addresses another subsystem.
D:UEFI firmware entries locate an EFI executable by device and path; its scope differs from recovering from a newly broken kernel image, so it cannot meet the stated requirement.
Question 602
df reports 70% free disk space, but creating a new file fails with No space left on device. Which resource might be exhausted?
- Create a supported filesystem before mounting
- NFS server or network path availability
- Available filesystem inodes
- Offline filesystem consistency check after backup
Explanation
Correct Answer
C:When all available inodes are allocated, new files cannot be created even if the block-space report shows free capacity. df -i distinguishes inode exhaustion from block capacity.
Incorrect Answers
A:A block device with no recognized filesystem cannot be mounted as ext4 or XFS until formatted appropriately; the administrator instead needs recognizing inode exhaustion despite available bytes; this option serves a different purpose.
B:A hard network mount can wait on unavailable remote I/O; it answers a neighboring question rather than recognizing inode exhaustion despite available bytes.
D:Repair tools may alter filesystem structures, so preserve data and unmount the affected filesystem first; this does not achieve recognizing inode exhaustion despite available bytes on the system described.
Question 603
Boot fails because an NVMe storage driver is unavailable before root is mounted, although it exists under /lib/modules. What needs rebuilding?
- /etc/fstab entry for /data
- Last kernel console messages and panic context
- Initramfs with the needed storage module
- Boot the prior known-good kernel
- UEFI firmware boot entry or EFI path
Explanation
Correct Answer
C:A driver only available on the inaccessible root filesystem arrives too late. Including it in the initramfs permits early root discovery.
Incorrect Answers
A:Incorrect identifiers, mount types or required mounts can push boot into emergency mode; using it would leave loading the boot storage module before root access unresolved.
B:Panic logs, call traces and boot parameters narrow the failing driver or subsystem; the requirement is loading the boot storage module before root access, outside the effect of this choice.
D:A working older kernel isolates the regression and restores service without immediately rebuilding disks; the evidence calls for loading the boot storage module before root access, not merely this separate operation.
E:UEFI firmware entries locate an EFI executable by device and path; the task involves loading the boot storage module before root access, whereas this addresses another subsystem.
Question 604
A startup script reports command not found only under systemd, while the same command works in an interactive shell. What is the likely cause?
- Different PATH and execution environment
- systemctl list-dependencies and failed units
- /etc/fstab entry for /data
- UEFI firmware boot entry or EFI path
Explanation
Correct Answer
A:Systemd services receive a controlled environment that may omit interactive profile additions. Using full executable paths avoids reliance on a login shell.
Incorrect Answers
B:Systemd dependency inspection reveals which required unit blocked target activation; that is not equivalent to resolving execution-path differences in systemd.
C:Incorrect identifiers, mount types or required mounts can push boot into emergency mode; the administrator instead needs resolving execution-path differences in systemd; this option serves a different purpose.
D:UEFI firmware entries locate an EFI executable by device and path; it answers a neighboring question rather than resolving execution-path differences in systemd.
Question 605
A service starts but fails repeatedly because its configured user does not exist. Which unit setting should be checked?
- User= identity in the service unit
- Linux out-of-memory killer
- Invalid process memory access
- Service shutdown handler and TimeoutStopSec
Explanation
Correct Answer
A:Systemd switches to the configured identity before starting a process. A missing or invalid user prevents expected execution.
Incorrect Answers
B:The OOM killer selects processes when memory reclaim cannot meet allocations under policy and cgroup limits; it will not provide fixing a service's nonexistent execution identity under these conditions.
C:A segmentation fault occurs when a process accesses memory in violation of protection rules; using it would leave fixing a service's nonexistent execution identity unresolved.
D:Slow cleanup or a stuck process can exceed systemd's grace period; the requirement is fixing a service's nonexistent execution identity, outside the effect of this choice.
Question 606
The name app.internal resolves to an IPv6 address but the service listens only on IPv4. Which mismatch causes the failed connection?
- PTR record in the reverse zone
- systemd-resolved stub listener
- Incomplete DNS resolution due to blocked TCP fallback
- AAAA resolution versus listener address family
Explanation
Correct Answer
D:A client may prefer an IPv6 address when a name has AAAA records. A service bound only to IPv4 cannot accept that IPv6 connection.
Incorrect Answers
A:Reverse DNS resolves addresses through PTR records under in-addr.arpa or ip6.arpa; the evidence calls for diagnosing an address-family mismatch after DNS lookup, not merely this separate operation.
B:On distributions using the stub resolver, 127.0.0.53 forwards local DNS requests to configured upstream servers; the task involves diagnosing an address-family mismatch after DNS lookup, whereas this addresses another subsystem.
C:DNS can use TCP for truncated or larger responses; its scope differs from diagnosing an address-family mismatch after DNS lookup, so it cannot meet the stated requirement.
Question 607
Traffic succeeds on an untagged switch port but VLAN-tagged guests cannot reach their subnet. What should be confirmed?
- Ethernet autonegotiation and cabling quality
- VLAN ID and switch trunk tagging agreement
- Physical-layer corruption or faulty link components
- End-to-end effective MTU across the VLAN path
Explanation
Correct Answer
B:Tagged frames must use matching VLAN IDs across the host, bridge and switch trunk. Untagged success does not validate that mapping.
Incorrect Answers
A:Poor cable pairs or mismatched negotiation can limit link speed; it answers a neighboring question rather than correcting mismatched VLAN tagging.
C:Frame check sequence errors are consistent with corrupted link-layer traffic; this does not achieve correcting mismatched VLAN tagging on the system described.
D:Encapsulation overhead and lower-device limits determine usable payload size; it will not provide correcting mismatched VLAN tagging under these conditions.
Question 608
Remote connections hang on a server even though ICMP ping works, and tcpdump shows SYN packets arriving with no SYN-ACK leaving. Which issue is plausible?
- Path MTU or MSS mismatch
- Packet loss and link errors on the affected path
- Local TCP filtering or listen-path failure
- Application backend availability and proxy health
Explanation
Correct Answer
C:Arriving SYNs without responses suggest host-side filtering or failure to service the target socket. Verify socket and firewall rules next.
Incorrect Answers
A:Oversized packets and blocked fragmentation feedback can stall larger transfers; the requirement is diagnosing inbound SYN packets without replies, outside the effect of this choice.
B:Retransmissions often reflect loss, congestion or reordering rather than application parsing; it does not deliver diagnosing inbound SYN packets without replies, which is decisive in this scenario.
D:HTTP 503 indicates that the request reached an HTTP-speaking server which cannot currently serve it; the task involves diagnosing inbound SYN packets without replies, whereas this addresses another subsystem.
Question 609
A packet capture shows repeated TCP RST from the destination immediately after SYN. What does that usually suggest?
- Overlap of allowed SSH host-key algorithms
- Destination explicitly rejecting the connection
- Packet loss and link errors on the affected path
- UDP-specific reachability and filtering test
- Application backend availability and proxy health
Explanation
Correct Answer
B:TCP RST often indicates a closed port or intentional rejection policy. It differs from silent packet drops producing timeouts.
Incorrect Answers
A:The client and server must negotiate a mutually supported acceptable host-key algorithm; that is not equivalent to interpreting a TCP reset during connection setup.
C:Retransmissions often reflect loss, congestion or reordering rather than application parsing; it answers a neighboring question rather than interpreting a TCP reset during connection setup.
D:ICMP success does not establish UDP port reachability; that changes a different part of the system; the requested outcome is interpreting a TCP reset during connection setup.
E:HTTP 503 indicates that the request reached an HTTP-speaking server which cannot currently serve it; this does not achieve interpreting a TCP reset during connection setup on the system described.
Question 610
Large HTTPS uploads hang while tiny requests succeed across a VPN. Which network property is a strong suspect?
- Local TCP filtering or listen-path failure
- Overlap of allowed SSH host-key algorithms
- Application backend availability and proxy health
- Path MTU or MSS mismatch
Explanation
Correct Answer
D:Oversized packets and blocked fragmentation feedback can stall larger transfers. Inspect MTU, MSS clamping and ICMP packet-too-big handling.
Incorrect Answers
A:Arriving SYNs without responses suggest host-side filtering or failure to service the target socket; it will not provide diagnosing size-dependent TCP stalls under these conditions.
B:The client and server must negotiate a mutually supported acceptable host-key algorithm; using it would leave diagnosing size-dependent TCP stalls unresolved.
C:HTTP 503 indicates that the request reached an HTTP-speaking server which cannot currently serve it; the requirement is diagnosing size-dependent TCP stalls, outside the effect of this choice.