WorksheetsMCQ on CO4 Disk Storage and File Systems
Total questions: 100
Worksheet time: 50mins
A file is updated 5 times in 5 minutes, and each update causes the entire segment (1 MB) to be rewritten. What total disk space is used after 5 updates?
1 MB
5 MB
500 KB
2 MB
To optimize performance, which of the following should be scheduled during low system load in LFS?
Checkpoint writing
Segment cleaning
Inode table compression
Log truncation
After a power failure, the LFS system appears to have lost recent file changes. Which component might not have updated in time?
Segment summary
Inode map
Checkpoint region
Directory cache
You observe that your LFS file system is using disk space inefficiently. What is the likely cause?
Infrequent checkpointing
Overwritten inodes
High number of partially full segments
File read bottlenecks
Why does an LFS require more RAM compared to traditional file systems?
It loads all file contents into RAM
It stores directory entries in memory
It caches inode maps and segment usage tables
It performs compression in RAM
You design a system using LFS but experience high memory usage. What is a practical solution?
Increase inode size
Cache only recent parts of the inode map
Remove the log structure
Use swap space
You need to reduce the garbage collection overhead in an LFS-based file system. Which approach would be most effective?
Decrease the segment size
Use a better segment cleaning policy
Increase the number of inodes
Enable journaling
Distributed file systems store data across multiple nodes to ensure availability. Which system is an example?
Ceph
NTFS
FAT32
ext3
In HDFS, data is divided into fixed-size blocks and distributed. What is the default block size?
64 MB
128 MB
256 MB
512 MB
Contiguous allocation stores files in consecutive blocks. What is the main drawback of this method?
Fast access
External fragmentation
Simplified bookkeeping
No fragmentation
Sequential allocation is best suited for which type of files?
Random access files
Sequential access files
Database files
Executable files
Indexed allocation uses an index block to store pointers. What advantage does it have over sequential allocation?
Supports direct access
Uses less disk space
Faster sequential access
No need for index blocks
Which type of allocation allows random access without external fragmentation?
Contiguous
Sequential
Indexed
Linked
Ceph uses an algorithm to distribute data evenly across disks. What is this algorithm called?
RAID
CRUSH
Consistent Hashing
Replication
In a distributed computing environment, systems often need access to shared data. NFS allows multiple machines to access files from a common location. Identify the key benefit provided by NFS.
Faster local storage access
Data redundancy
Centralized file access over a network
Improved CPU performance
High-speed data access is crucial for real-time applications. Some storage technologies are designed for ultra-low latency. Select the technology optimized for fastest access.
SATA SSD
NVMe SSD
HDD
Tape storage
Operating systems provide a consistent way for applications to interact with files, regardless of underlying storage types. This logical layer abstracts file operations. Identify this component.
File system interface
I/O scheduler
Disk controller
Virtual memory
Organizations using AWS need scalable, shared file storage across multiple instances. One service in AWS supports elastic file systems with POSIX compliance. Choose the correct AWS service.
Google Filestore
Azure Blob Storage
Amazon EFS
Dropbox
Unix systems maintain metadata like ownership, timestamps, and access permissions in a special structure. This structure does not store file data directly. Identify the structure used.
Directory entry
Inode
File descriptor
Superblock
When creating a file in Unix, the system must reserve a metadata structure for it. A specific system call handles this allocation. Choose the call responsible for inode reservation.
alloc
namei
ialloc
ifree
Cloud-based applications require scalable storage for images, logs, and unstructured data. Microsoft Azure provides an object storage solution optimized for this. Select the suitable service.
Azure Files
Blob Storage
Google Cloud Storage
Amazon Glacier
Disk scheduling manages order of disk I/O requests.
FCFS
Round Robin
LRU
SSTF
Block device drivers handle:
Byte-by-byte data
Fixed-size blocks
Network packets
Interrupt requests
SCAN scheduling moves disk head:
Randomly
In one direction only
Back and forth
Only forward
The elevator algorithm in disk scheduling is also known as:
SSTF
SCAN
C-SCAN
FIFO
Character device drivers provide access to:
Random access files
Sequential data streams
Block data
System logs
Device drivers act as an interface between:
Hardware and applications
Kernel and user
Hardware and kernel
User and file system
AI-enhanced disk scheduling improves efficiency by:
Predicting requests
Increasing buffer size
Encrypting data
Reducing disk size
Kernel-bypass I/O frameworks like io_uring use shared queues to manage multiple I/O requests without kernel switches. Which data structure manages these requests?
Submission & completion queues
Blocking calls
Polling
Synchronous I/O
RDMA allows direct memory access between computers bypassing the OS. How does RDMA reduce CPU overhead?
Kernel bypass
More interrupts
Kernel threads
Buffer copying
DPDK achieves high throughput by avoiding kernel networking stack. Which technique does it use?
Poll mode drivers
Interrupt-driven I/O
Kernel modules
File buffering
RDMA requires specific hardware for zero-copy transfers. Which hardware enables this?
RDMA-enabled NICs
Standard NICs
CPUs
Disk controllers
DPDK processes packets in user space. How are packets received from NIC hardware?
Poll mode drivers
Interrupt-driven I/O
Kernel modules
Buffers
RDMA uses certain transport protocols for data transfer. Which is commonly used?
InfiniBand
TCP
UDP
HTTP
Kernel-bypass I/O like io_uring reduces latency by minimizing what?
Context switches
Memory usage
Disk writes
File descriptor access
A system using LFS writes 5 small files in rapid succession. What performance benefit can be expected compared to a traditional file system?
Lower memory usage
Reduced read latency
Improved sequential write throughput
Increased file fragmentation
An LFS writes a new version of a file every minute. Each version takes 4 KB. How many versions can fit in a 1 MB segment before the segment becomes full?
64
256
128
512
A flash-based embedded system uses LFS. You notice system slowdowns during heavy writes. What would be the most effective optimization?
Increase segment size
Run garbage collection during idle times
Disable segment cleaning
Decrease checkpoint interval
If each inode is 128 bytes and you need to manage 10,000 files, how much space is needed for just the inodes in LFS?
1.28 MB
128 KB
10 MB
512 KB
An LFS has 10 segments with the following live data percentages: [90, 10, 20, 80, 40, 30, 15, 95, 60, 50]. Which segments should the cleaner target first to reclaim space efficiently?
Segments with > 50% live data
Segments with < 50% live data
Segments with 100% live data
Segments in sequential order
You are evaluating LFS and ext4 for a mobile OS. The workload is 90% writes and runs on NAND flash. Which system offers better long-term performance and endurance?
ext4
LFS
Both are equal
Depends on RAM size
In indexed allocation, what data structure is crucial for fast access to file blocks?
Index block
File descriptor
Directory entry
Inode
Analyze fragmentation: Which allocation type suffers the most from internal fragmentation?
Contiguous
Indexed
Sequential
None
What is the effect of replication factor in distributed file systems like Ceph and HDFS?
Improves performance
Reduces storage usage
Increases fault tolerance
Lowers latency
In distributed file systems, which layer typically handles client requests for file operations?
Metadata server
Data node
Client node
Network interface
Evaluate the impact of contiguous allocation on file deletion and creation speed.
Faster
Slower
No effect
Depends on file size
Ceph uses a distributed hash table. What is the primary purpose of this structure?
Data encryption
Data location and lookup
File compression
Data replication
Which allocation method is most space-efficient for small files with random access?
Contiguous
Sequential
Indexed
Linked
Network file access can be achieved through several protocols. One major benefit of NFS compared to FTP is its design to maintain no client state. Identify this feature.
Secure login
Stateless communication
File compression
Authentication
Operating systems may support different file systems like ext4, FAT32, or NTFS. One abstraction layer allows uniform access across them. Identify this layer.
Implements disk I/O
Manages CPU registers
Provides abstraction over multiple FS
Handles process scheduling
Data-intensive applications such as machine learning benefit from a particular memory type that combines speed and persistence. Identify the suitable storage medium.
HDD
Optical Disk
Persistent Memory
Tape
A fully managed service in Google Cloud Platform provides scalable file storage compatible with NFS. It is used with VMs and containers. Choose the correct service.
Azure Files
Google Filestore
Dropbox
Box
Buffer cache in Unix stores frequently accessed file blocks in memory to improve system performance.
Speeds up file system access
Increases CPU speed
Reduces network traffic
Caches DNS queries
Object storage manages data as discrete units with metadata. What key advantage does it have over block storage?
Better metadata handling
Lower cost
Higher read/write speed
More complex file system
Driver buffering improves performance by:
Reducing CPU cycles
Minimizing disk seeks
Avoiding deadlocks
Lowering power usage
The main limitation of SSTF is:
Starvation
Long wait times
High CPU usage
Excessive buffering
Which disk scheduling algorithm guarantees fairness by processing requests in arrival order?
SSTF
FCFS
SCAN
C-SCAN
Device drivers provide hardware abstraction by:
Hiding hardware details
Increasing latency
Direct memory access
Network management
Block device drivers handle I/O in:
Bytes
Fixed-size blocks
Variable-sized packets
Packets
Batching I/O in io_uring affects syscall overhead by?
Reducing it
Increasing it
Causing latency
No change
Which RDMA protocol supports lossless Ethernet?
RoCE
TCP
UDP
HTTP
Kernel bypassing in DPDK impacts security by requiring?
User space security
Same kernel protections
Reduced security
No change
io_uring signals I/O completion through?
Completion queue polling
Blocking waits
Interrupts
Signals
Hardware ensures RDMA transfers do not interfere with system by?
Memory protection
Software locks
Kernel mediation
Polling
Binding NICs to DPDK drivers causes?
NIC unavailable to kernel stack
NIC available everywhere
System instability
No effect
io_uring instance is created with which syscall?
io_uring_setup()
epoll()
select()
read()
A Log-Structured File System writes data in segments of 16 MB each. If a disk has a write bandwidth of 200 MB/s and the cleaning efficiency is 80%, how much time will it take to write 1 GB of new data including cleaning overhead?
5.0 seconds
6.25 seconds
8.0 seconds
10.0 seconds
If a file system using Copy-on-Write (CoW) has a disk bandwidth of 300 MB/s and snapshot creation takes 0.5 s, what is the maximum size of the snapshot data that can be created without impacting ongoing writes assuming writes take 3 seconds for 900 MB?
100 MB
150 MB
200 MB
250 MB
In an HDFS cluster with 5 data nodes, each with 4 TB storage, and a replication factor of 3, what is the effective usable storage?
6.67 TB
12.5 TB
20 TB
60 TB
A file system uses contiguous allocation on a disk of 1 TB with block size 4 KB. If a file is 100 MB, how many contiguous blocks does it occupy?
25600 blocks
25000 blocks
26000 blocks
20000 blocks
A disk with 1000 cylinders is at cylinder 200, and requests arrive for cylinders 100, 350, 600, and 900. Using SCAN (elevator) scheduling towards higher cylinders first, calculate total head movement.
1200 cylinders
1100 cylinders
1300 cylinders
No Match
An RDMA device transfers data at 10 Gbps. How many megabytes per second does this correspond to?
1250 MB/s
1000 MB/s
1024 MB/s
800 MB/s
A Log-Structured File System segment size is 32 MB. Cleaning reclaims 70% of the segment. If writing a 1 GB file requires cleaning 5 segments, how much extra write overhead (in MB) does cleaning cause?
48 MB
96 MB
112 MB
160 MB
If a Btrfs snapshot copies 20% of a 500 MB file system, how much additional storage is needed?
100 MB
200 MB
400 MB
20 MB
A Ceph cluster uses replication factor 3 with 15 TB raw storage. What is the effective storage?
10 TB
5TB
45TB
15TB
If a file system uses single-level index blocks of 4 KB, block pointers 4 bytes, how many data blocks can it point to?
4096
1024
2048
512
Maximum File Size (Direct + Single Indirect + Double Indirect) Given: 12 direct blocks, 1 single indirect block, 1 double indirect block, block size 4 KB, pointer size 4 bytes. Calculate max file size.
64 MB
260 MB
4 GB
16 MB
Cache size 64 buffers of 4 KB. A 1 MB file is accessed repeatedly 5 times. How many cache hits after first read?
960
2560
1024
1280
NVMe device has 15 µs latency and throughput 350,000 IOPS. What is the max throughput in MB/s (block size 4 KB)?
1.4 GB/s
1.3 GB/s
1.2 GB/s
1.1 GB/s
Azure charges $0.0184 per GB per month. Using 750 GB for 4 months, what's the cost?
$55.20
$60.30
$45.20
$50.00
Google Filestore reports 10,000 IOPS on a 1 TB instance. If block size is 8 KB, what is throughput in MB/s?
80 MB/s
78 MB/s
90 MB/s
85 MB/s
Using FCFS, Requests at cylinders: 45, 180, 10, 90, current at 50, calculate total head movement.
275
300
330
295
Using SSTF, Requests at cylinders: 45, 180, 10, 90, current at 50, calculate total head movement.
300
400
330
No Match
If AI reduces average seek time by 20% from 8 ms to what?
6.0 ms
7.0 ms
5.5 ms
6.4 ms
Block device with block size 512 bytes does 1000 IOPS, throughput?
2 MB/s
0.5 MB/s
0.6 MB/s
1 MB/s
Character device transmits 115,200 baud with 8 data bits, 1 stop bit, no parity. Effective data rate in bytes per second?
12,800 B/s
14,400 B/s
11,520 B/s
10,000 B/s
If io_uring reduces latency by 30% from 50 µs, what is new latency?
30 µs
40 µs
38 µs
35 µs
10 Gbps link with 80% utilization, effective bandwidth MB/s?
800 MB/s
1250 MB/s
1000 MB/s
1024 MB/s
If ialloc takes 0.2 ms and ifree 0.1 ms, time for 500 allocations and 300 frees?
140 ms
150 ms
130 ms
160 ms
If NFS adds 5 ms latency per operation and local FS latency is 1 ms, total for 100 ops?
700 ms
400 ms
500 ms
600 ms
Persistent memory write bandwidth 3 GB/s, file size 600 MB, write time?
0.6 s
0.2 s
0.3 s
0.5 s
Disk with queue depth 32 and avg service time 10 ms, max IOPS?
2500
3000
3200
2800
AWS EFS throughput 100 MB/s per TB, with 3 TB used, throughput?
250 MB/s
200 MB/s
350 MB/s
300 MB/s
If disk I/O wait is 5 ms and CPU runs at 2 GHz with 1 instruction per cycle, how many instructions lost during wait?
10 million
5 million
7 million
15 million
Doubling block size from 4 KB to 8 KB decreases number of blocks by?
100%
75%
50%
25%
If DPDK transfers 8 million packets/sec at avg packet size 512 bytes, throughput?
3 GB/s
4 GB/s
5 GB/s
6 GB/s
A Unix file system allocates 1 inode per 4 KB. If a disk has 40 GB usable space, how many inodes can be created?
20,480
10,240
102,400
40,960
A file uses 10 direct blocks, 1 single indirect block (holds 256 addresses), and 1 double indirect ( 2562 ). What is max file size with 4 KB blocks?
268 MB
300 MB
264 MB
270 MB
A disk request queue is optimized by AI-based scheduling, reducing average seek time from 9 ms to 4 ms. If 5000 operations are processed, how much time is saved?
22 seconds
20 seconds
30 seconds
25 seconds
Azure Blob storage charges 0.02perGB/monthand 0.05 per 1000 transactions. You store 500 GB and perform 100,000 operations/month. What's total cost?
$10.50
$15
$11
$13
