CHAPTER 3

FILE SYSTEMS & DISK

MANAGEMENT
Presented by FRU NGANG
PLAN
• 3.1 File systems
• 3.1.1 Unix FSH
• 3.1.2 Mounting a File system
• 3.1.3 Some Unix file systems
• 3.1.4 The /proc FS
• 3.2 Disk Management
• 3.2.1 Disk interfaces and controllers
• 3.2.2 RAID
• 3.2.3 Swap space
• 3.2.4 Disk storage
• 3.2.5 Logical Volume management
• 3.3 Backing up the system
3.1 UNIX
• File system refers to the files and directories stored on a computer.
• A file system can have different formats called file system types.
These formats determine how the information is stored as files and
directories.
• Some file system types store redundant copies of the data, while
some file system types make hard drive access faster.
• The term file system has all of the following meanings:
• A logical collection of files, directories, named pipes, links, and other UNIX
items and metadata that are arranged in a hierarchy.
• A particular instance of a logical collection of these items that are arranged
in a hierarchy. They might be on local or remote disks or in computer
memory.
• A program that is designed to provide the functions and data of one type of
file system.
3.1.1 The Unix File
Hierarchy
• The FS structure is the most basic of organization in an OS.
• File systems break files down into two logical categories:
• Shareable vs. unshareable files
• Variable vs. static files
• Shareable files are those that can be accessed locally
and by remote hosts;
• unshareable files are only available locally.
• Variable files, such as documents, can be changed at
any time;
• static files, such as binaries, do not change without
an action from the system administrator.
3.1.1 The Unix File
Hierarchy
 3.1.1 The Unix File
System Hierarchy
• /boot/ directory contains static files required for the system
to boot. These files are essential for the system to boot
properly.
• /dev/ contains device nodes. These nodes could be actual
devices attached to the system or virtual devices provided
by the kernel. The udev daemon creates or removes the
devices in /dev/.
• /etc/ is used for system configuration and should not contain
any binaries.
• /proc/ a virtual file system that contains information about
hardware devices and processes derived from the kernel.
 3.1.1 The Unix File
System Hierarchy
• /lib/ should contain only libraries needed to execute binaries
in /bin and /sbin. These are important for booting the system
and executing commands within the root file system.
• /media/ contains subdirectories used as mountpoints for
removable devices such as usb storage and DVDs, CDs etc.
• /mnt/ is reserved for temporarily mounted file systems such as
NFS system mounts.
• /sbin/ contains binaries essential for systems administration.
• /sys/ contains files for hotplug devices as used by the sysfs.
 3.1.1 The Unix File
System Hierarchy
• /usr/ directory is for files that can be shared
across multiple machines. The /usr/
directory is often on its own partition and
mounted read only.
• The /var directory contains variable
data written to it by the system.
3.1.2 Mounting a file system
• file systems on different partitions and removable
devices like CDs, DVDs, or USB flash drives can be
attached to a certain point (that is, the mount point)
in the directory tree, and detached again.
• To attach or detach a file system, use the mount or
unmount commands.
• To view all currently attached file systems use
mount with no options

•
# mount
• Options of the mount command
3.1.3 Some Unix FS types
• Read the functionalities and limitations of
•
Ext2
•
Ext3
•
Ext4
•
XFS
•
Btrfs
•
Jfs
•
Reiserfs
3.3 Disk Controllers and Interfaces
Few popular standards
IDE (integrated device electronics)
ATA (AT attachment interface)
SCSI (small computer systems interface)
Fibre Channel
Differences
Performance
Parallelism

1
3.3.1 SCSCI
Many versions
 SCSI-1 (1986) 8-bits, 5MB/s
 SCSI-2 (1990) added command
queuing, DMA, more
 Fast SCSI-2 8-bits, 10MB/s
 Fast/wide SCSI-2 16-bits, 20MB/s
 Ultra SCSI 8 bits, 20MB/s
 Wide Ultra SCSI 16bits,40MB/s
 Wide Ultra2 SCSI 16bits, 80MB/s
 Wide Ultra3 SCSI 16bits, 160MB/s
 Ultra-320, Ultra-640 SCSI

1
 3.3.2 IDE a.k.a ATA
Integrated Drive Electronics / AT Attachment
Very short cable lengths (18in!)
ATA-2 added DMA and LBA
(get beyond BIOS 504MB limit)
ATA-3 added power management, self-monitoring (16MB/s)
Ultra-ATA added Ultra DMA/33, /66, and /133 modes (33-
133MB/s)
Hard disks with this interface were last produced in 2013
ATAPI interface allows non-ATA devices to connect like CD-ROMS

1
 3.3.2 SATA
Now standard equipment
Fast: 150-600MB/s (16GBit/s now available)
Software compatible with parallel ATA
One drive per controller
Thin cables

1
 3.3.3 SCSCI vs
SCSI traditionally beats SATA technically, but may not be worth
the price premium
 In single-user systems, SATA will provide 85%, cheaply
For best possible performance, SCSI is often better
 e.g., in servers and multi-user systems
 handles multiple simultaneous reqs + more devices better
 higher-end equipment (faster, better warranty, etc.)
SATA technology is quite good
 Usually better price/performance than SCSI
 Still subject to much debate

1
 3.3.4 Controller

A controller card is a device that sits between a host system and the storage
network and allows the two to communicate with each other. Host adapters
can be integrated in the motherboard or be on a separate expansion card

There are two types of controller cards: Host bus Adapters and RAID
Controller Cards.

HBA is an expansion card that plugs into a slot (PCI-e) and provides fast
reliable non-RAID I/O between the host and storage devices.

HBA is low cost, high connectivity, limited functionality, best performance

RAID is similar to HBA but adds redundancy, optimize performance reduce
latency.

1
 3.3.4 Controller cards
 RAID (Redundant Array of Independent Disks) is a data storage
structure that allows a system to combine two or more physical storage
devices into a logical unit that is seen by the attached system as a single
drive.
 RAID can be hardware based or software implemented.
 Hardware RAID resides on PCI-X or PCI-e, or on a motherboard – integrated
RAID-on-Chip (ROC).
 Better performance than software RAID, RAID
cards can easily be swapped out for
replacement and upgrades. Data can be backed up to prevent loss in a power failure.
 More expensive than software Raid
 Software Raid runs entirely on the CPU of the host system
 Lower cost
 Lower RAID performance as
CPU also powers the OS and Apps, no data backup

1
 3.3.5
Different levels of RAID implementation.
RAID 0 also known as striped volume. Its spreads data over
multiple drives to enhance performance There is no
redundancy and hence no data protection. It provides
highest performance

1
 3.3.5
RAID 1 creates an exact copy (mirror) of some data or a disk.
This layout is useful when read performance or reliability is more
important than write performance or the resulting data storage
capacity.

1
 3.3.5
RAID 2 stripes data at the bit level (rather than block level)
and uses Hamming codes for error correction.
A separate disk is used for parity.
It is rarely used in practice

2
 3.3.5
RAID 3 stripes data at the byte level with a dedicated parity
disk.
It is rarely used in practice

2
3.3.5
RAID 4 stripes data at the byte level with a
dedicated parity disk.
It is rarely used in practice

2
 3.3.5
RAID5 Combines data striping for enhanced performance with
distributed parity for data protection to provide a recovery path in
case of failure.
Best cost/performance balance for multi-drive environment.

2
 3.3.5
RAID 6 Provides double redundancy and the ability to sustain
two drive failures. Data is striped across at least 4 physical
drives. A second parity scheme is used to store and recover
data.

2
 3.3.5 RAID Combinations
RAID 10 Combines RAID 0 (data striping) and RAID 1 (disk
mirroring).Ithas the highest performance with highest data
protection.
RAID 50 combines multiple RAID 5 sets with data striping
(RAID 0) to increase capacity and performance without
adding disks to each RAID 5 array. Increased capacity and
performance for multi array RAID5 environments.
RAID 60 Combines multiple RAID 6 sets with data
striping (RAID 0) to increase capacity and performance
without adding disks to each RAID 6 array. It has the
highest Fault tolerance and the highest data protection.

2
 3.3.5 Configuring RAID Linu
1. Check your version of linux (#uname –r)
2. Check to see if your system contains the /proc/mdstat file

ls /proc/mdstat
3. Use the modprobe command to check if you can load the
different RAID kernel modules

modprobe raid6
4. Check to see if your system has the Multiple Disk or Device
Administration (mdadm) utility installed.

dpkg –s mdadm

2
 3.3.5 Configuring RAID Linu
5. Getting a drive ready for RAID membership
•
Partition the disk first using fdisk or parted
•
Use the lsblk command to view the drives on the system
•
Implement the raid array
•
#mdadm
#mdadm [--mode] raid-device [options]
component-devices
6) Create the RAID membership
# mdadm -C /dev/md0 -l 6 -n 4 \
> /dev/sdf1 /dev/sdg1 /dev/sdh1 /dev/sdi1

2
 3.4 Adding a disk to Linux
 Install new hardware
 verify that hardware is recognized by BIOS or controller
 – Boot, make certain device files already exist in /dev e.g., /dev/sdc
 Use fdisk/parted (or similar) to partition the drive
 Verify the system type on each partition

Use mke2fs (-t ext4) on each regular partition
 To create (an ext4) filesystem
 Use mkswap to initialize swap partitions

Add entries to /etc/fstab

Mount by hand, then reboot to verify everything

2
hdparm: test/set hd params
hdparm will do simple performance tests
sudo /sbin/hdparm –Tt /dev/sda
Read man hdparm

2
 3.4 Disk
 A partition is a logical subset of the physical disk, and information about
partitions are stored in a partition table. This table includes information
about the first and last sectors of the partition and its type, and further
details on each partition.

 Usually each partition is seen by an operating system as a separate “disk”,

even if they all reside in the same physical media.

 On Linux, each partition is assigned to a directory under /dev like

/dev/sda1 or /dev/sda2

3
 3.4 Disk

Drives are divided into one or more partitions that are treated
independently

Partitions make backups easier, confine damage

Typically have at least two or three
 root partition (one)
 everything needed to bring system up in single-user mode
(often copied onto another disk for emergencies)
 swap partition (at least one)
stores virtual memory when physical memory is insufficient
 user partition(s)
home directories, data files, etc.
 boot partition - boot loader, kernel, etc.

3
 3.4 Disk partitions (MBR and
 There are two main ways of storing partition information on
hard disks. The first one is MBR (Master Boot Record), and
the second one is GPT (GUID Partition Table).
 MBR

The partition table is stored on the first sector of a disk, called
the Boot Sector, along with a boot loader, which on Linux
systems is usually the GRUB bootloader.

But MBR has a series of limitations that hinder its use on
modern systems, like the inability to address disks of more
than 2 TB in size, and the limit of only 4 primary partitions per
disk.

3
 3.4 Disk partitions (MBR and
 GUID Partition Table (GPT)

There is no practical limit on disk size, and the maximum
number of partitions are limited only by the operating system
itself.

It is more commonly found on more modern machines that use
UEFI instead of the old PC BIOS.

3
 3.4 Disk partitions (MBR and
Managing MBR partitions using fdisk
The standard utility for managing MBR
partitions on Linux is fdisk. This is an
interactive, menu-driven utility. To use it,
type fdisk followed by the device name
corresponding to the disk you want to edit
Ex sudo fdisk /dev/sda
You can create, edit or delete partitions at will,
but nothing will be written to disk unless you
use the write (w) command and q to exit fdisk

3
 3.4 Disk partitions (MBR and
Managing MBR partitions using fdisk
fdisk commands
p: to print the partition table
n : to create a partition
F: to show unallocated space
d: to delete a partition
t: change partition type ( swap, etc)

3
 3.4 Disk partitions (MBR and
Managing GUID partitions using gdisk
gdisk commands
p: to print the partition table
n : to create a partition
F: to show unallocated space
d: to delete a partition
s: to sort the partitions on a disk to
avoid gaps in the numbering
sequence

3
 3.5 Tools to manage partitions
Tools for manipulating partitions are:
fdisk and its derivatives like cfdisk, sfdisk
parted and its variants like qtparted and gparted

3
 3.5 Logical Volumes
Partitions are static and sometimes you would want to change
them
LVM (Linux Logical Volume Manager) lets you combine partitions
and drives to present an aggregate volume as a regular block
device (just like a disk or partition)
 Use and allocate storage more efficiently
 Move logical volumes among different physical devices
 Grow and shrink logical volume sizes dynamically
 Take “snapshots” of whole filesystems
 Replace on-line drives without interrupting service
Similar systems are available for other OSes

3
3.6
Sample Organization

sda1 sdc2 Physical volume

diskvg volume group

usrlv rootlv varlv logical

3
3.6

volumes

4
 3.6
The configuration file of LVM is found in the directory /etc/lvm.
The file /etc/lvm/lvm.conf contains the global parameters.

The configuration of different volumes (physical, groups and

logical) are not found in a file but in a structure present in the
device itself.

The fdsik command displays the partition table of a disk

 # fdisk -l sda

4
 3.7 Backup
Assignment:
Read the following backup types:
Full
Incremental
Differential
Snapshot

4
 3.7 Backup
GUI and/or web-based solutions
Amanda
Bacula
Bareos
Duplicity
BAckupPC

4
 3.7 Backup
Command line utilities
cpio
dd
dump/restore
rsync
star
tar