Red Hat Enterprise Linux 7 Implementation

File System Creation and Mounting

Module Topics

Mounting File Systems Manually
Unmounting File Systems
Accessing Removable Storage Devices
Disk Partitioning
Managing MBR Partitions with fdisk
Managing GPT Partitions with gdisk
Creating File Systems
Mounting File Systems
Swap Space Concepts
Create a Swap Space
Activate a Swap Space

Mounting File Systems Manually

Need to manually mount file system on SATA/PATA or SCSI device
Use mount
- First argument specifies file system to mount
- Second argument specifies directory where file system resides after mounting
  - Called mount point
Two ways mount expects file system argument:
- Device file of partition holding file system
  - Resides in /dev
- UUID
  - Universal unique identifier of file system
  - Does not change unless file system is re-created, even if device order changes or new devices are added to system

Transcript:

A file system residing on a SATA/PATA or SCSI device needs to be mounted manually to access it. The mount command allows the root user to manually mount a file system. The first argument of the mount command specifies the file system to mount. The second argument specifies the target directory where the file system is made available after mounting it. The target directory is referred to as a mount point.

The mount command expects the file system argument in one of two different ways:

The device file of the partition holding the file system, residing in /dev.
The UUID, a universal unique identifier of the file system.

As long as a file system is not recreated, the UUID stays the same. The device file can change; for example, if the order of the devices is changed or if additional devices are added to the system.

Mounting File Systems Manually

Partitions and UUID

blkid gives overview of:
- Partitions with file system on them
- File system UUID
- File system used to format partition
  [root@server1~]# blkid /dev/vda1:UUID="46f543fd-78c9-4526-a857-244811be2d88" TYPE="xfs"
Can mount file system on existing directory
- Default /mnt directory provides entry point for mount points
- Recommended: Create subdirectory under /mnt to use as mount point

To mount by partition device file:

[root@server1~]# mount /dev/vdb1 /mnt/mydata

To mount by file system UUID:

[root@server1~]# mount UUID="46f543fd-78c9-4526-a857-244811be2d88" /mnt/mydata

If mount point is not empty, file in mount point not accessible while file system is mounted
All files written to mount point directory appear on file system mounted there

Transcript:

The blkid command gives an overview of existing partitions with a file system on them and the UUID of the file system, as well as the file system used to format the partition.

A file system can be mounted on an existing directory. The /mnt directory exists by default and provides an entry point for mount points. It is used for manually mounting disks. It is recommended to create a subdirectory under /mnt and use that subdirectory as a mount point unless there is a reason to mount the file system in another specific location in the file system hierarchy.

Mount by device file of the partition that holds the file system.

Mount the file system by universal unique id, or the UUID, of the file system.

If the directory acting as mount point is not empty, the files that exists in that directory are not accessible as long as a file system is mounted there. All files written to the mount point directory end up on the file system mounted there.

Unmounting File Systems

To unmount file system, use umount

Expects mount point as argument

Example:

Change to /mnt/mydata directory
Try to unmount device on /mnt/mydata mount point

It fails

[root@server1~]# cd /mnt/mydata
[root@server1mydata]# umount /mnt/mydata
umount:/mnt/mydata: target is busy.
        (In some cases useful info about processes that use
         the device is found by lsof(8) or fuser(1))

Unmounting File Systems

Processes That Prevent Unmounting

Cannot unmount if process accesses mount point
Need to stop access to unmount

To list all open files and process accessing them in directory, use lsof

Use to identify processes preventing file system from unmounting

[root@server1mydata]# lsof /mnt/mydata
COMMAND  PID USER   FD   TYPE DEVICE SIZE/OFF NODE NAME
bash    1593 root  cwd    DIR  253,2        6  128 /mnt/mydata
lsof    2532 root  cwd    DIR  253,2       19  128 /mnt/mydata
lsof    2533 root  cwd    DIR  253,2       19  128 /mnt/mydata

After identifying, can take action
- Wait for process to complete
- Send SIGTERM or SIGKILL signal
- Change directory to directory outside mount point
  [root@server1mydata]# cd [root@server1~]# umount /mnt/mydata
Common reason mount point is busy - Current directory of shell prompt is below active mount point
- bash process is accessing mount point
- To unmount device, change to directory outside mount point

Transcript:

Unmounting is not possible if the mount point is accessed by a process. For umount to be successful, the process needs to stop accessing the mount point.

The lsof command lists all open files and the process accessing them in the provided directory. It is useful to identify which processes currently prevent the file system from successful unmounting.

Once the processes are identified, an action can be taken, such as waiting for the process to complete or sending a SIGTERM or SIGKILL signal to the process. In this case, it is sufficient to change the current working directory to a directory outside the mount point.

A common cause for the file system on the mount point to be busy is if the current working directory of a shell prompt is below the active mount point. The process accessing the mount point is bash. Changing to a directory outside the mount point allows the device to be unmounted.

Accessing Removable Storage Devices

Graphical desktop environment automatically mounts removable media, i.e., USB flash devices and drives
Mount point for removable medium: /run/media/<user>/<label>
- <user> is user logged in to graphical environment
- <label> is name given to file system when created

Must unmount USB media before removing it
Removing without unmounting can cause data loss

References

Man pages: mount(8), umount(8), and lsof(8)

Disk Partitioning

Disk partitioning: Lets you divide hard drive into multiple logical storage units (partitions)
Can use different partitions to perform different functions, such as:
- Limit available space to applications or users
- Allow multibooting of different operating systems from same disk
- Separate operating system and program files from user files
- Create separate area for OS virtual memory swapping
- Limit disk space usage to improve diagnostic/imaging performance

Disk Partitioning

MBR Partitioning Scheme

Master Boot Record (MBR): Dictates how to partition disks on BIOS firmware systems
Supports maximum four primary partitions
On Linux, can use extended and logical partitions to create maximum 15 partitions
Partition size data stored as 32-bit values
Disks partitioned with MBR have maximum 2 TiB disk/partition size limit
2 TiB limit presents real-world problem encountered frequently in production environments
MBR scheme being superseded by GUID Partition Table (GPT)

Transcript:

Since 1982, the Master Boot Record (MBR) partitioning scheme has dictated how disks should be partitioned on systems running BIOS firmware. This scheme supports a maximum of four primary partitions. On Linux systems, with the use of extended and logical partitions, administrator can create a maximum of 15 partitions. Since partition size data are stored as 32-bit values, disks partitioned with the MBR scheme have a maximum disk and partition size limit of 2 TiB.

With the advent of hard drives with ever-increasing capacity, the 2 TiB disk and partition size limit of the aged MBR partitioning scheme is no longer a theoretical limit, but rather a real-world problem that is being encountered more and more frequently in production environments. As a result, the legacy MBR scheme is in the process of being superseded by the new GUID Partition Table (GPT) for disk partitioning.

Disk Partitioning

GPT Partitioning Scheme

GPT: Standard for laying out partition tables on hard disks for Unified Extensible Firmware Interface (UEFI) firmware systems
Part of UEFI standard
Addresses many limitations of MBR-based scheme
Supports up to 128 partitions
Allocates 64 bits for logical block addresses
Can accommodate partitions and disks of up to 8 ZiB (8 billion TiB)
- 8 ZiB limit based on 512-byte block size
- With 4,096-byte blocks, limitation increases to 64 ZiB
Uses 128-bit GUIDs to identify disks and partitions
Offers redundancy of partition table information
- Primary GPT resides at head of disk
- Backup copy/secondary GPT housed at end of disk
Employs CRC checksum to detect errors and corruption

Transcript:

For systems running Unified Extensible Firmware Interface (UEFI) firmware, GPT is the standard for laying out partition tables on physical hard disks. GPT is part of the UEFI standard and addresses many of the limitations imposed by the old MBR-based scheme. Per UEFI specifications, GPT defaults to supporting up to 128 partitions. Unlike MBR, which uses 32 bits for storing logical block addresses and size information, GPT allocates 64 bits for logical block addresses. This allows GPT to accommodate partitions and disks of up to 8 zebibyte (ZiB), or 8 billion tebibytes.

GPT’s 8-ZiB limit is based on a 512-byte block size. With hard drive vendors transitioning to 4,096-byte blocks, this limitation will increase to 64 ZiB.

In addition to addressing the limitations of the MBR partitioning scheme, GPT also offers some additional features and benefits. Per its namesake, GPT uses 128-bit GUIDs to uniquely identify each disk and partition. In contrast to MBR, which has a single point of failure, GPT offers redundancy of its partition table information. The primary GPT resides at the head of the disk, while a backup copy, the secondary GPT, is housed at the end of the disk. In addition, GPT employs the use of CRC checksum to detect errors and corruption in the GPT header and partition table.

Managing MBR Partitions with `fdisk`

Partition editors let administrators make changes to disk partitions
- Creating partitions, deleting partitions, changing partition types
Can use fdisk partition editor to perform operations for MBR scheme disks

Managing MBR Partitions with `fdisk`

Creating MBR Disk Partitions

Specify disk device on which to create partition:

[root@server1~]# fdisk /dev/vdb
Welcome to fdisk (util-linux 2.23.2).

Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.

Command (m for help):

Request new primary or extended partition:
```
Partition type:
   p   primary (0 primary, 0 extended, 4 free)
   e   extended
Select (default p):p
```
- If you need more than four partitions, create three primary and one extended partition; use extended partition as container for multiple logical partitions

Managing MBR Partitions with `fdisk`

Specify partition number:
```
Partition number (1-4, default 1):1
```
Specify first disk sector on which new partition will start:
```
First sector (2048-20971519, default 2048):2048
```

Specify last disk sector on which new partition will end:

Last sector, +sectors or +size{K,M,G} (6144-20971519, default 20971519):1050623

Can also enter number representing desired partition size in sectors:

Last sector, +sectors or +size{K,M,G} (6144-20971519, default 20971519):+52488

Also can specify partition size in units KiB, MiB, or GiB:

Last sector, +sectors or +size{K,M,G} (6144-20971519, default 20971519):+512M

After you enter partition’s ending boundary, fdisk displays partition creation confirmation:
```
Partition 1 of type Linux and of size 512 MiB is set
```

Transcript:

Specify partition number.
This partition number serves as the identification number of the new partition on the disk for use in future partition operations. The default value is the lowest unused partition number.
Specify the first sector on the disk that the new partition will start on.
The default value is the first available sector on the disk.
Specify the last sector on the disk that the new partition will end on.
The default value is the last of the available, unallocated sectors contiguous to the new partition’s first sector.

In addition to the ending sector number, fdisk can also accept a number representing the desired size of the partition expressed in sectors.

The final, and most user-friendly, input option offered by fdisk is to specify the size of the new partition in units of KiB, MiB, or GiB.

Once the partition’s ending boundary is entered, fdisk will then display a confirmation of the partition creation.

Managing MBR Partitions with `fdisk`

Define partition type:

To change partition type to anything other than Linux, use t

To display table of hex codes for all partition types, use L

Command (m for help):t
Selected partition 1
Hex code (type L to list all codes):82
Changed type of partition 'Linux' to 'Linux swap / Solaris'

Save partition table changes:

Command (m for help):w
The partition table has been altered!

Calling ioctl() to re-read partition table.

WARNING: Re-reading the partition table failed with error 16: Device or resource busy.
The kernel still uses the old table. The new table will be used at
the next reboot or after you run partprobe(8) or kpartx(8)
Syncing disks.

Initiate kernel re-read of new partition table:
```
[root@server1~]# partprobe /dev/vdb
```
- Must use w to write partition table edits to disk
- To discard erroneous commands, exit fdisk without running w

Transcript:

Define partition type.
If the newly created partition should have a type other than Linux, enter the t command to change a partition’s type. Enter the hex code for the new partition type. If needed, a table of the hex codes for all partition types can be displayed with the L command. Setting the partition type correctly is crucial, since some tools rely on it to function properly. For example, when the Linux kernel encounters a partition of type 0xfd, Linux RAID, it will attempt to autostart the RAID volume.
Save partition table changes.
Issue the w command to finalize the partition creation request by writing the changes to the disk’s partition table and exiting the fdisk program.

Initiate a kernel re-read of the new partition table.

Run the partprobe command with the disk device name as an argument to force a re-read of its partition table.

The fdisk program queues all partition table edits and writes them to disk only when the administrator issues the w command to write all partition table changes to disk. If the w command is not executed prior to exiting the interactive fdisk session, all requested changes to the partition table will be discarded and the disk’s partition table will remain unchanged. This feature is especially useful when erroneous commands are issued to fdisk. To discard the erroneous commands and avoid their unintended consequences, simply exit fdisk without saving the partition table changes.

Managing MBR Partitions with `fdisk`

Removing MBR Disk Partitions

Specify disk containing partition to remove:

[root@server1~]# fdisk /dev/vdb
Welcome to fdisk (util-linux 2.23.2).

Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.

Command (m for help):

Identify partition number of partition to delete:

Command (m for help):p

Disk /dev/vdb:10.7 GB, 10737418240 bytes, 20971520 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical):512 bytes / 512 bytes
I/O size (minimum/optimal):512 bytes / 512 bytes
Disk label type:dos
Disk identifier:0xd2368130

   Device Boot      Start         End      Blocks   Id  System
/dev/vdb1            2048     1050623      524288   82  Linux swap / Solaris

Managing MBR Partitions with `fdisk`

Request partition deletion:

Command (m for help):d
Selected partition 1
Partition 1 is deleted

Save partition table changes:

Command (m for help):w
The partition table has been altered!

Calling ioctl() to re-read partition table.

WARNING:Re-reading the partition table failed with error 16: Device or resource busy.
The kernel still uses the old table. The new table will be used at
the next reboot or after you run partprobe(8) or kpartx(8)
Syncing disks.

Initiate kernel re-read of new partition table:
```
[root@server1~]# partprobe /dev/vdb
```

Managing GPT Partitions with `gdisk`

Can use gdisk partition editor to manage partitions for GPT partitioning scheme disks
Use gdisk with GPT partitioning scheme
- GPT support for fdisk experimental

Managing GPT Partitions with `gdisk`

Creating GPT Disk Partitions

Specify disk device on which to create partition:

[root@server1~]# gdisk /dev/vdb
GPT fdisk (gdisk) version 0.8.6

Partition table scan:
  MBR:not present
  BSD:not present
  APM:not present
  GPT:not present

Creating new GPT entries.

Command (? for help):

Request new partition:
```
Command (? for help):n
```
Specify partition number:
```
Partition number (1-128, default 1):1
```

Managing GPT Partitions with `gdisk`

Specify disk location from which new partition will start:
- Option 1: Enter absolute disk sector number representing tfirst sector of new partition
  First sector (34-20971486, default = 2048) or {+-}size{KMGTP}:2048
- Option 2: Indicates partition’s starting sector by position relative to first or last sector of first contiguous block of free sectors on disk
  - Specify input in KiB, MiB, GiB, TiB, or PiB
  - Example: **+**512M signifies sector position 512 MiB after beginning of next group of contiguous available sectors
  - **-**512M denotes sector position 512 MiB before end of group of contiguous available sectors
Specify last disk sector on which new partition will end:
```
Last sector (2048-20971486, default = 20971486) or {+-}size{KMGTP}:1050623
```
- Can specify end boundary of new partition in KiB, MiB, GiB, TiB, or PiB from beginning or end of group of contiguous available sectors
  - Example: **+**512M signifies ending partition position 512 MiB after first sector
    Last sector (2048-20971486, default = 20971486) or {+-}size{KMGTP}:+512M
  - **-**512M indicates ending partition position 512 MiB before the end contiguous available sectors
    Last sector (2048-20971486, default = 20971486) or {+-}size{KMGTP}:-512M

Transcript:

Specify the disk location that the new partition will start from.
gdisk allows for two different input types. The first input type is an absolute disk sector number representing the first sector of the new partition.

The second input type indicates the partition’s starting sector by its position relative to the first or last sector of the first contiguous block of free sectors on the disk. Using this relative sector position format, input is specified in units of KiB, MiB, GiB, TiB, or PiB.

For example, a value of +512M signifies a sector position that is 12 MiB after the beginning of the next group of contiguous available sectors. On the other hand, a value of -512M denotes a sector positioned 512 MiB before the end of this group of contiguous available sectors.
Specify the last sector on the disk that the new partition will end on.
The default value is the last of the available, unallocated sectors contiguous to the new partition’s first sector.

In addition to the absolute ending sector number, gdisk also offers the more user-friendly input option of specifying the end boundary of the new partition in units of KiB, MiB, GiB, TiB, or PiB from the beginning or end of the group of contiguous available sectors. A value of +512M signifies an ending partition position that is 512 MiB after the first sector.

A value of -512M indicates an ending partition position that is 512 MiB before the end of the contiguous available sectors.

Managing GPT Partitions with `gdisk`

Define partition type:
- Use hex code to change partition type to anything other than Linux
- Use L to display table of hex codes for all partition types
  Current type is 'Linux filesystem' Hex code or GUID (L to show codes, Enter = 8300):8e00 Changed type of partition to 'Linux LVM'

Save partition table changes:

Command (? for help):w

Final checks complete. About to write GPT data. THIS WILL OVERWRITE EXISTING
PARTITIONS!!

Do you want to proceed? (Y/N):y
OK; writing new GUID partition table (GPT) to /dev/vdb.
The operation has completed successfully.

Initiate kernel re-read of new partition table:
```
[root@server1~]# partprobe /dev/vdb
```
- Must use w to write partition table edits to disk
- To discard erroneous commands, exit gdisk without running w

Transcript:

Define partition type.
New partitions created by gdisk default to type Linux file system. If a different partition type is desired, enter the corresponding hex code. If needed, a table of the hex codes for all partition types can be displayed with the L command.
Save partition table changes.
Issue the w command to finalize the partition creation request by writing the changes to the disk’s partition table. Enter y when gdisk prompts for a final confirmation.
Initiate a kernel re-read of the new partition table.
Run the partprobe command with the disk device name as an argument to force a re-read of its partition table.

The gdisk program queues all partition table edits and writes them to disk only when the administrator issues the w command to write all partition table changes to disk. If the w command is not executed prior to exiting the interactive gdisk session, all requested changes to the partition table will be discarded and the disk’s partition table will remain unchanged. This feature is especially useful when erroneous commands are issued to gdisk. To discard the erroneous commands and avoid their unintended consequences, simply exit gdisk without saving the partition table changes.

Managing GPT Partitions with `gdisk`

Removing GPT Disk Partitions

Specify disk containing partition to remove:

[root@server1~]# gdisk /dev/vdb
GPT fdisk (gdisk) version 0.8.6

Partition table scan:
  MBR:protective
  BSD:not present
  APM:not present
  GPT:present

Found valid GPT with protective MBR; using GPT.

Command (? for help):

Identify partition number of partition to delete:

Command (? for help):p
Disk /dev/vdb:20971520 sectors, 10.0 GiB
Logical sector size:512 bytes
Disk identifier (GUID):8B181B97-5259-4C8F-8825-1A973B8FA553
Partition table holds up to 128 entries
First usable sector is 34, last usable sector is 20971486
Partitions will be aligned on 2048-sector boundaries
Total free space is 19922877 sectors (9.5 GiB)

Number  Start (sector)    End (sector)  Size       Code  Name
   1            2048         1050623   512.0 MiB   8E00  Linux LVM

Managing GPT Partitions with `gdisk`

Request partition deletion:
```
Command (? for help):d
Using 1
```

Save partition table changes:

Command (? for help):w

Final checks complete. About to write GPT data. THIS WILL OVERWRITE EXISTING
PARTITIONS!!

Do you want to proceed? (Y/N):y
OK; writing new GUID partition table (GPT) to /dev/vdb.
The operation has completed successfully.

Initiate kernel re-read of new partition table:
```
[root@server1~]# partprobe /dev/vdb
```

Creating File Systems

After creating block device, must apply file system format to it
- File system applies structure to block device
- Lets you store and retrieve data device
Red Hat Enterprise Linux supports many file system types
- Two common types: xfs and ext4
- anaconda (Red Hat Enterprise Linux installer) uses xfs by default

To apply file system to block device, use mkfs

To specify file system type, use -t

If no type specified, ext2 used by default

[root@server1~]# mkfs -t xfs /dev/vdb1
meta-data=/dev/vdb1              isize=256    agcount=4, agsize=16384 blks
         =                       sectsz=512   attr=2, projid32bit=1
         =                       crc=0
data     =                       bsize=4096   blocks=65536, imaxpct=25
         =                       sunit=0      swidth=0 blks
naming   =version 2              bsize=4096   ascii-ci=0 ftype=0
log      =internal log           bsize=4096   blocks=853, version=2
         =                       sectsz=512   sunit=0 blks, lazy-count=1
realtime =none                   extsz=4096   blocks=0, rtextents=0

Mounting File Systems

After applying file system format, need to attach file system into directory structure
Lets user space utilities access or write files on device

Mounting File Systems

Manually Mounting File Systems

To manually attach device to directory location (mount point), use mount
Specify:
- Device
- Mount point
- Any option desired to customize device behavior
  [root@server1~]# mount /dev/vdb1 /mnt

Can also use mount to view:

Currently mounted file systems
Mount points

Options

[root@server1~]# mount | grep vdb1
/dev/vdb1 on /mnt type xfs (rw,relatime,seclabel,attr2,inode64,noquota)

Manually mounting file system lets you verify that formatted device is accessible/working as desired
After system reboot, file system exists and has intact data but is not mounted into directory tree again
To permanently mount file system, add listing for file system to /etc/fstab

Mounting File Systems

Persistently Mounting File Systems

To configure device to be mounted to mount point at system boot, add device listing in /etc/fstab

/etc/fstab is white space-delimited file with six fields per line

[root@server1~]# cat /etc/fstab
#
# /etc/fstab
# Created by anaconda on Thu Mar 20 14:2:46 2014
#
# Accessible filesystems, by reference, are maintained under '/dev/disk'
# See man pages fstab(5), findfs(8), mount(8) and/or blkid(8) for more info
#
UUID=7a20315d-ed8b-4e75-a5b6-24ff9e1f9838  /  xfs  defaults  1 1

First field specifies device to be used (example uses UUID)
- Could also use device file; for example, /dev/vdb1
  - UUID is stored in file system superblock and created at same time as file system
  - Recommended: Use UUID; remains intact if block device identifiers change

Mounting File Systems

blkid

To scan block devices connected to machine, use blkid
- Also reports on data like assigned UUID and file system format
  [root@server1~]# blkid /dev/vdb1 /dev/vdb1:UUID="226a7c4f-e309-4cb3-9e76-6ef972dd8600" TYPE="xfs"
Second field is mount point at which to attach device into directory hierarchy
- Mount point should already exist
- If not, use mkdir to create
Third field contains file system type applied to block device
Fourth field lists options to apply to mounted device to customize behavior
- Required field
- defaults contains set of commonly used options
- For other options, see mount man page
Last two fields are dump flag and fsck order
- Use dump flag with dump to make a backup of device contents
- fsck order field determines if fsck should run at boot time if file system did not unmount cleanly
- fsck order value indicates order in which file systems should have fsck run on them
  UUID=226a7c4f-e309-4cb3-9e76-6ef972dd8600 /mnt xfs defaults 1 2

Transcript:

The blkid command can be used to scan the block devices connected to a machine and report on data like the assigned UUID and file system format.

The second field is the mount point where the device should be attached into the directory hierarchy. The mount point should already exist; if it does not, it can be created with mkdir.

The third field contains the file system type that has been applied to the block device.

The fourth field is the list of options that should be applied to the device when mounted to customize the behavior. This field is required, and there is a set of commonly used options called defaults. Other options are documented in the mount man page.

The last two fields are the dump flag and fsck order. The dump flag is used with the dump command to make a backup of the contents of the device. The fsck order field determines if the fsck should be run at boot time, in the event that the file system was not unmounted cleanly. The value of the fsck order indicates the order in which file systems should have fsck run on them if multiple file systems are required to be checked.

Mounting File Systems

Incorrect entry in /etc/fstab may render the machine unbootable
To verify entry is valid:
- Unmount new file system
- Mount file system back into place using mount -a, which reads /etc/fstab
  - If mount -a returns error, correct this before rebooting machine

References

Man pages: fdisk(8), gdisk(8), mkfs(8), mount(8), fstab(5)

Swap Space Concepts

Swap space: Disk area you can use with Linux kernel memory management subsystem
- Use to supplement system RAM by holding inactive pages of memory
- System RAM + swap spaces = Virtual memory
When system memory usage exceeds limit, kernel:
- Combs RAM looking for idle memory pages assigned to processes
- Writes idle page to swap area
- Reassigns RAM page to another process
If program requires access to page written to disk, kernel:
- Locates another idle memory page
- Writes it to disk
- Recalls needed page from swap area
Swap slow compared to RAM
Keep use of swap to a minimum

Transcript:

A swap space is an area of a disk which can be used with the Linux kernel memory management subsystem. Swap spaces are used to supplement the system RAM by holding inactive pages of memory. The combined system RAM plus swap spaces is called virtual memory.

When the memory usage on a system exceeds a defined limit, the kernel will comb through RAM looking for idle memory pages assigned to processes. The kernel writes the idle page to the swap area, and will reassign the RAM page to be used by another process. If a program requires access to a page that has been written to disk, the kernel will locate another idle page of memory, write it to disk, then recall the needed page from the swap area.

Since swap areas reside on disk, swap is incredibly slow when compared with RAM. While it is used to augment system RAM, usage of swap spaces should be kept to a minimum whenever possible.

Create a Swap Space

Create partition
Set partition type as 82 Linux Swap
Format swap signature on device

Create a Swap Space

Create a Partition

Use tool such as fdisk to create partition

Example: Create 256 MiB partition

[root@server1~]# fdisk /dev/vdb
Welcome to fdisk (util-linux 2.23.2).

Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.

Device does not contain a recognized partition table
Building a new DOS disklabel with disk identifier 0x34e4e6d7.

Command (m for help):n
Partition type:
   p   primary (0 primary, 0 extended, 4 free)
   e   extended
Select (default p):p
Partition number (1-4, default 1):1
First sector (2048-20971519, default 2048):Enter
Using default value 2048
Last sector, +sectors or +size{K,M,G} (2048-20971519, default 20971519):+256M
Partition 1 of type Linux and of size 256 MiB is set

Command (m for help):p

Disk /dev/vdb:10.7 GB, 10737418240 bytes, 20971520 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical):512 bytes / 512 bytes
I/O size (minimum/optimal):512 bytes / 512 bytes
Disk label type:dos
Disk identifier:0x34e4e6d7

   Device Boot      Start         End      Blocks   Id  System
/dev/vdb1            2048      526335      262144   83  Linux

Create a Swap Space

Assign the Partition Type

Recommended: Change created partition’s type (system ID) to 82 Linux Swap

Setting partition type helps administrators determine partition’s purpose

Command (m for help):t
Selected partition 1
Hex code (type L to list all codes):82
Changed type of partition 'Linux' to 'Linux swap / Solaris'

Command (m for help):p

Disk /dev/vdb:10.7 GB, 10737418240 bytes, 20971520 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical):512 bytes / 512 bytes
I/O size (minimum/optimal):512 bytes / 512 bytes
Disk label type:dos
Disk identifier:0x34e4e6d7

   Device Boot      Start         End      Blocks   Id  System
/dev/vdb1            2048      526335      262144   82  Linux swap / Solaris

Create a Swap Space

Format the Device

To apply swap signature to device, use mkswap

Writes single block of data at beginning of device

Leaves rest of device unformatted

[root@server1~]# mkswap /dev/vdb1
Setting up swapspace version 1, size = 262140 KiB
no label, UUID=fbd7fa60-b781-44a8-961b-37ac3ef572bf

Activate a Swap Space

To activate formatted swap space, use swapon
To activate all swap spaces listed in /etc/fstab, use swapon -a

Can call swapon on device

[root@server1~]# free
             total       used       free     shared    buffers     cached
Mem:      1885252     791812    1093440      17092        688     292024
-/+ buffers/cache:    499100    1386152
Swap:           0          0          0
[root@server1~]# swapon /dev/vdb1
[root@server1~]# free
             total       used       free     shared    buffers     cached
Mem:      1885252     792116    1093136      17092        692     292096
-/+ buffers/cache:    499328    1385924
Swap:      262140          0     262140

Activate a Swap Space

Persistently Activate Swap Space

Swap space likely needs to automatically activate with every machine boot
- Must configure this in /etc/fstab
Can use swapoff to deactivate swap space
Works only if swapped data can be written to other active swap spaces or back into memory

Activate a Swap Space

Add a Previously Created Swap Space

UUID=fbd7fa60-b781-44a8-961b-37ac3ef572bf  swap  swap  defaults  0 0

Field	Description
1	`UUID` or raw device name
2	Swap devices use `swap` as placeholder value for mount point
3	File system type for swap space is `swap`
4	Options, such as `defaults` which includes `auto` mount option to automatically activate swap space at boot
5	`dump` flag; backing up not required for swap spaces
6	`fsck` order; file system checking not required for swap spaces

Transcript:

The following is an example line in /etc/fstab adding a previously created swap space.

The example uses the UUID as the first field. The UUID is stored in the swap signature stored on the device, and was part of the output of mkswap. If the output of mkswap has been lost, the blkid command can be used to scan the system and report on all attached block devices. If the administrator does not wish to use the UUID, the raw device name can also be used in the first field.

The second field is typically reserved for the mount point. However, for swap devices, which are not accessible through the directory structure, this field is the placeholder value swap.

The third field is the file system type. The file system type for a swap space is swap.

The fourth field is for options. In the example, the option defaults is used. defaults includes the mount option auto, which is what causes the swap space to be automatically activated at boot.

The final two fields are the dump flag and fsck order. Swap spaces require neither backing up nor file system checking.

Activate a Swap Space

Swap Space Priorities

Default: Swap spaces used in series
- First activated swap space used until full, then kernel starts using second swap space
To display swap space priorities, use swapon -s
To set priorities, use pri= mount option
If swap spaces have same priority, kernel writes to them round-robin

References

Man pages: mkswap(8), swapon(8), swapoff(8), mount(8), fdisk(8)

Summary

Mounting File Systems Manually
Unmounting File Systems
Accessing Removable Storage Devices
Disk Partitioning
Managing MBR Partitions with fdisk
Managing GPT Partitions with gdisk
Creating File Systems
Mounting File Systems
Swap Space Concepts
Create a Swap Space
Activate a Swap Space

Module Completion

Nice job!

Click the button below to complete this module of the course:

Red Hat Enterprise Linux 7 Implementation

File System Creation and Mounting

Module Topics

Mounting File Systems Manually

Mounting File Systems Manually

Unmounting File Systems

Unmounting File Systems

Accessing Removable Storage Devices

Disk Partitioning

Disk Partitioning

Disk Partitioning

Managing MBR Partitions with fdisk

Managing MBR Partitions with fdisk

Managing MBR Partitions with fdisk

Managing MBR Partitions with fdisk

Managing MBR Partitions with fdisk

Managing MBR Partitions with fdisk

Managing GPT Partitions with gdisk

Managing GPT Partitions with gdisk

Managing GPT Partitions with gdisk

Managing GPT Partitions with gdisk

Managing GPT Partitions with gdisk

Managing GPT Partitions with gdisk

Creating File Systems

Mounting File Systems

Mounting File Systems

Mounting File Systems

Mounting File Systems

Mounting File Systems

Swap Space Concepts

Create a Swap Space

Create a Swap Space

Create a Swap Space

Create a Swap Space

Activate a Swap Space

Activate a Swap Space

Activate a Swap Space

Activate a Swap Space

Summary

Module Completion

Managing MBR Partitions with `fdisk`

Managing MBR Partitions with `fdisk`

Managing MBR Partitions with `fdisk`

Managing MBR Partitions with `fdisk`

Managing MBR Partitions with `fdisk`

Managing MBR Partitions with `fdisk`

Managing GPT Partitions with `gdisk`

Managing GPT Partitions with `gdisk`

Managing GPT Partitions with `gdisk`

Managing GPT Partitions with `gdisk`

Managing GPT Partitions with `gdisk`

Managing GPT Partitions with `gdisk`