file systems

Question 1

main function of a file system

Answer 1

main permanent data storage

Question 2

is capacity a problem nowadays?

Answer 2

no, 2TB even for PC

Question 3

if capacity is not a problem what is

Answer 3

backup becoming a problem

Question 4

logical view of file systems

Answer 4

tree structure of files
together with read/write operation and creation of directories

Question 5

physical view of file system

Answer 5

sequence of blocks, which can be read and
written. OS has to map logical view to physical view, must impose tree structure and assign blocks for each file

Question 6

what are the 2 main possibilities to realise filesystems

Answer 6

linked lists
index allocation

Question 7

describe linked list in file systems

Answer 7

Each block contains pointer to next
⇒ Problem: random access (seek()) costly: have to go
through whole file until desired position

Question 8

describe index allocation in file systems

Answer 8

Store pointers in one location: so-called
index block (similar to page table). To cope with vastly
differing file sizes, may introduce indirect index blocks

Question 9

what are index blocks in unix

Answer 9

inodes

Question 10

what do is something additional inodes do

Answer 10

they store additional info about the file
eg size,
permissions)

Question 11

what does FAT stand for

Answer 11

File Allocation Table

Question 12

describe FAT

Answer 12

F(ile) A(llocation) T(able) – dates back to 70s.
Useful for explaining filesystem concepts, modern filesystems
are more complicated
Variants FAT12. FAT16, FAT32 define number of bits per
FAT entry – we focus on FAT16
Sector = disk unit (e.g. 512 byte), aka block
Cluster = multiple sectors (factor 1, 2, 4, . . . , 128)
(here: assume cluster = 1 sector)
Uses linked list (“cluster chain”) to group clusters

Question 13

what does sector mean in FAT

Answer 13

Sector = disk unit (e.g. 512 byte), aka block

Question 14

what does Cluster mean In FAT

Answer 14

Cluster = multiple sectors (factor 1, 2, 4, . . . , 128)
(here: assume cluster = 1 sector)

Question 15

what is use to grop clusters in FAT

Answer 15

linked lists

Question 16

what is the max volume size of FAT16

Answer 16

2 GB (216 · 32 kB)

Question 17

what is the max file size of FAT16

Answer 17

2GB

Question 18

what is the max number of files of FAT16

Answer 18

65,460 (32kb clusters)

Question 19

how do new filesystems overcome FAT limits

Answer 19

overcome these limits,
using other data structures (e.g. B-tree for dir structure,
bitmap for allocation)

Question 20

describe cachine

Answer 20

Disk blocks used for storing directories or recently used files cached in main memory
Blocks periodically written to disk
⇒ Big effiency gain
Inconsistency arises when system crashes
Reason why computers must be shutdown properly

Question 21

describe journaling file systems

Answer 21

Define Transaction points: Points where cache is written to disk
⇒ Have consistent state
Keep log-file for each write-operation
Log enough information to unravel any changes done after latest transaction point

Question 22

disk access contains 3 parts
what are they?

Answer 22

Seek
latency
transfer

Question 23

what is seek in data access

Answer 23

head moves to appropriate track

Question 24

what is latency in disk access

Answer 24

correct block is underhead

Question 25

what is transfer in disk access

Answer 25

data transfer

Question 26

describe disk access time in HDDS

Answer 26

HDDs: Time necessary for seek and latency dwarfs transfer time ⇒ Distribution of data and scheduling algorithms have vital impact on performance for HDDs, less so for SSDs

Question 27

name 2 disk acheduling algorithms

Answer 27

FCFS Shortest seek time first SCAN scheduling

Question 28

describe FCFS in disk scheduling

Answer 28

easiest to implement, but: may require lots of head movements

Question 29

describe shortest seek time first in disk scheduling

Answer 29

select job with minimal head movement

Question 30

problems with shortest seek time first

Answer 30

may cause starvation tracks in the middle of disks preferred.

Question 31

Describe scan scheduling in disk scheduling

Answer 31

Head continuously scans the disk from en to end (lift strategy) ⇒ solves the fairness and starvation problem of SSTF

Question 32

improvement to scan scheduling

Answer 32

look scheduling

Question 33

what does look scheduing do?

Answer 33

head only moved as far as last request (lift strategy)

Question 34

describe disk algorithm used for swap space management

Answer 34

Particular tasks may require different disk access algorithms Example : Swap space management Speed absolutely crucial ⇒ different treatment: Swap space stored on separate partition Indirect access methods not used Special algorithms used for access of blocks Optimised for speed at the cost of space (eg increased internal fragmentation

Question 35

describe linux implementation

Answer 35

Interoperability with Windows and Mac requires support of different file systems (eg vfat) ⇒ Linux implements common interface for all filesystems Common interface called virtual file system virtual file system maintains inodes for files and directories caches, in particular for directories superblocks for file systems All system calls (eg open, read, write and close) first go to virtual file system If necessary, virtual file system selects appropriate operation from real file system

Question 36

describe kernel disk scheduler

Answer 36

File systems Disk Scheduler Kernel makes it possible to have different schedulers for different file systems Default scheduler (Completely Fair Queuing) based on lift strategy have in addition separate queue for disk requests for each process queues served in Round-Robin fashion Have in addition No-op scheduler: implements FIFO Suitable for SSD’s where access time for all sectors is equal