Difference between revisions of "Fstab"

From ArchWiki
Jump to: navigation, search
m (Avoid citing upstream unsupported kernel.)
(tmpfs: Make it clear that there is no default /tmp mount in /etc/fstab anymore because systemd mounts it by default as such.)
Line 177: Line 177:
 
[[Wikipedia:Tmpfs|tmpfs]] is a temporary filesystem that resides in memory and/or your swap partition(s), depending on how much you fill it up. Mounting directories as tmpfs can be an effective way of speeding up accesses to their files, or to ensure that their contents are automatically cleared upon reboot.
 
[[Wikipedia:Tmpfs|tmpfs]] is a temporary filesystem that resides in memory and/or your swap partition(s), depending on how much you fill it up. Mounting directories as tmpfs can be an effective way of speeding up accesses to their files, or to ensure that their contents are automatically cleared upon reboot.
  
Some directories where tmpfs is commonly used are [http://www.pathname.com/fhs/2.2/fhs-3.15.html /tmp], [http://www.pathname.com/fhs/2.2/fhs-5.9.html /var/lock] and [http://www.pathname.com/fhs/2.2/fhs-5.13.html /var/run]. Do NOT use it on [http://www.pathname.com/fhs/2.2/fhs-5.15.html /var/tmp], because that folder is meant for temporary files that are preserved across reboots. Arch uses a tmpfs {{ic|/run}} directory, with {{ic|/var/run}} and {{ic|/var/lock}} simply existing as symlinks for compatibility. It is also used for {{ic|/tmp}} in the default {{ic|/etc/fstab}}.
+
Some directories where tmpfs is commonly used are [http://www.pathname.com/fhs/2.2/fhs-3.15.html /tmp], [http://www.pathname.com/fhs/2.2/fhs-5.9.html /var/lock] and [http://www.pathname.com/fhs/2.2/fhs-5.13.html /var/run]. Do NOT use it on [http://www.pathname.com/fhs/2.2/fhs-5.15.html /var/tmp], because that folder is meant for temporary files that are preserved across reboots. Arch uses a tmpfs {{ic|/run}} directory, with {{ic|/var/run}} and {{ic|/var/lock}} simply existing as symlinks for compatibility. It is also used for {{ic|/tmp}} by the default systemd setup and does not require an entry in {{ic|/etc/fstab}} unless a specific configuration is needed.
  
 
{{Note|When using [[systemd]], temporary files in tmpfs directories can be recreated at boot by using [[Systemd#Temporary_files|tmpfiles.d]].}}
 
{{Note|When using [[systemd]], temporary files in tmpfs directories can be recreated at boot by using [[Systemd#Temporary_files|tmpfiles.d]].}}
Line 183: Line 183:
 
By default, a tmpfs partition has its maximum size set to half your total RAM, but this can be customized. Note that the actual memory/swap consumption depends on how much you fill it up, as tmpfs partitions do not consume any memory until it is actually needed.
 
By default, a tmpfs partition has its maximum size set to half your total RAM, but this can be customized. Note that the actual memory/swap consumption depends on how much you fill it up, as tmpfs partitions do not consume any memory until it is actually needed.
  
To use tmpfs for {{ic|/tmp}}, add this line to {{ic|/etc/fstab}}:
+
To explicitly set a maximum size, in this example to override the default {{ic|/tmp}} mount, use the {{ic|size}} mount option:
 
+
{{hc|/etc/fstab|
+
tmpfs  /tmp        tmpfs  nodev,nosuid                  0  0}}
+
 
+
You may or may not want to specify the size here, but you should leave the {{ic|mode}} option alone in these cases to ensure that they have the correct permissions (1777). In the example above, {{ic|/tmp}} will be set to use up to half of your total RAM. To explicitly set a maximum size, use the {{ic|size}} mount option:
+
  
 
{{hc|/etc/fstab|2=
 
{{hc|/etc/fstab|2=

Revision as of 18:45, 6 February 2014

The /etc/fstab file can be used to define how disk partitions, various other block devices, or remote filesystems should be mounted into the filesystem.

Each filesystem is described in a separate line. These definitions will be converted into systemd mount units dynamically at boot, and when the configuration of the system manager is reloaded. The default setup will automatically fsck and mount filesystems before starting services that need them to be mounted. For example, systemd automatically makes sure that remote filesystem mounts like NFS or Samba are only started after the network has been set up. Therefore, local and remote filesystem mounts specified in /etc/fstab should work out of the box. See man 5 systemd.mount for details.

The mount command will use fstab, if just one of either directory or device is given, to fill in the value for the other parameter. When doing so, mount options which are listed in fstab will also be used.

File example

A simple /etc/fstab, using kernel name descriptors:

/etc/fstab
# <file system>        <dir>         <type>    <options>             <dump> <pass>
/dev/sda1              /             ext4      defaults,noatime      0      1
/dev/sda2              none          swap      defaults              0      0
/dev/sda3              /home         ext4      defaults,noatime      0      2

Field definitions

The /etc/fstab file contains the following fields separated by a space or tab:

 <file system>        <dir>         <type>    <options>             <dump> <pass>
  • <file system> - the partition or storage device to be mounted.
  • <dir> - the mountpoint where <file system> is mounted to.
  • <type> - the file system type of the partition or storage device to be mounted. Many different file systems are supported: ext2, ext3, ext4, btrfs, reiserfs, xfs, jfs, smbfs, iso9660, vfat, ntfs, swap and auto. The auto type lets the mount command guess what type of file system is used. This is useful for optical media (CD/DVD).
  • <options> - mount options of the filesystem to be used. Note that some mount options are filesystem specific. Some of the most common options are:
  • auto - Mount automatically at boot, or when the command mount -a is issued.
  • noauto - Mount only when you tell it to.
  • exec - Allow execution of binaries on the filesystem.
  • noexec - Disallow execution of binaries on the filesystem.
  • ro - Mount the filesystem read-only.
  • rw - Mount the filesystem read-write.
  • user - Allow any user to mount the filesystem. This automatically implies noexec, nosuid, nodev, unless overridden.
  • users - Allow any user in the users group to mount the filesystem.
  • nouser - Allow only root to mount the filesystem.
  • owner - Allow the owner of device to mount.
  • sync - I/O should be done synchronously.
  • async - I/O should be done asynchronously.
  • dev - Interpret block special devices on the filesystem.
  • nodev - Don't interpret block special devices on the filesystem.
  • suid - Allow the operation of suid, and sgid bits. They are mostly used to allow users on a computer system to execute binary executables with temporarily elevated privileges in order to perform a specific task.
  • nosuid - Block the operation of suid, and sgid bits.
  • noatime - Don't update inode access times on the filesystem. Can help performance (see atime options).
  • nodiratime - Do not update directory inode access times on the filesystem. Can help performance (see atime options).
  • relatime - Update inode access times relative to modify or change time. Access time is only updated if the previous access time was earlier than the current modify or change time. (Similar to noatime, but doesn't break mutt or other applications that need to know if a file has been read since the last time it was modified.) Can help performance (see atime options).
  • discard - Issue TRIM commands to the underlying block device when blocks are freed. Recommended to use if the filesystem is located on an SSD.
  • flush - The vfat option to flush data more often, thus making copy dialogs or progress bars to stay up until all data is written.
  • nofail - Mount device when present but ignore if absent. This prevents errors being reported at boot for removable media.
  • defaults - the default mount options for the filesystem to be used. The default options for ext4 are: rw, suid, dev, exec, auto, nouser, async.
  • <dump> - used by the dump utility to decide when to make a backup. Dump checks the entry and uses the number to decide if a file system should be backed up. Possible entries are 0 and 1. If 0, dump will ignore the file system; if 1, dump will make a backup. Most users will not have dump installed, so they should put 0 for the <dump> entry.
  • <pass> - used by fsck to decide which order filesystems are to be checked. Possible entries are 0, 1 and 2. The root file system should have the highest priority 1 (unless its type is btrfs, in which case this field should be 0) - all other file systems you want to have checked should have a 2. File systems with a value 0 will not be checked by the fsck utility.

Identifying filesystems

There are three ways to identify a partition or storage device in /etc/fstab: by its kernel name descriptor, label or UUID. The advantage of using UUIDs or labels is that they are not dependent on the order in which the drives are (physically) connected to the machine. This is useful if the storage device order in the BIOS is changed, or if you switch the storage device cabling. Also, the BIOS may occasionally change the order of storage devices. Read more about this in the Persistent block device naming article.

To list basic information about the partitions, run:

$ lsblk -f
NAME   FSTYPE LABEL      UUID                                 MOUNTPOINT
sda                                                           
├─sda1 ext4   Arch_Linux 978e3e81-8048-4ae1-8a06-aa727458e8ff /
├─sda2 ntfs   Windows    6C1093E61093B594                     
└─sda3 ext4   Storage    f838b24e-3a66-4d02-86f4-a2e73e454336 /media/Storage
sdb                                                           
├─sdb1 ntfs   Games      9E68F00568EFD9D3                     
└─sdb2 ext4   Backup     14d50a6c-e083-42f2-b9c4-bc8bae38d274 /media/Backup
sdc                                                           
└─sdc1 vfat   Camera     47FA-4071                            /media/Camera

Kernel name

Run lsblk -f to list the partitions, and prefix them with /dev.

See the example.

Label

Note: Each label should be unique, to prevent any possible conflicts.

For detailed information on how to label a device or partition, see Persistent block device naming#by-label. Renaming the root partition has to be done from a "live" Linux distribution because the partition needs to be unmounted first.

Run lsblk -f to list the partitions, and prefix them with LABEL= :

/etc/fstab
# <file system>        <dir>         <type>    <options>             <dump> <pass> 
LABEL=Arch_Linux       /             ext4      defaults,noatime      0      1
LABEL=Arch_Swap        none          swap      defaults              0      0

UUID

All partitions and devices have a unique UUID. They are generated by filesystem utilities (e.g. mkfs.*) when you create or format a partition. See Persistent block device naming#by-uuid for details.

Run lsblk -f to list the partitions, and prefix them with UUID= :

Tip: If you would like to return just the UUID of a specific partition:
$ lsblk -no UUID /dev/sda2
/etc/fstab
# <file system>                            <dir>     <type>    <options>             <dump> <pass>
UUID=24f28fc6-717e-4bcd-a5f7-32b959024e26  /         ext4      defaults,noatime      0      1
UUID=03ec5dd3-45c0-4f95-a363-61ff321a09ff  /home     ext4      defaults,noatime      0      2
UUID=4209c845-f495-4c43-8a03-5363dd433153  none      swap      defaults              0      0

Tips and tricks

Automount with systemd

If you have a large /home partition, it might be better to allow services that do not depend on /home to start while /home is checked by fsck. This can be achieved by adding the following options to the /etc/fstab entry of your /home partition:

noauto,x-systemd.automount

This will fsck and mount /home when it is first accessed, and the kernel will buffer all file access to /home until it is ready.

Note: This will make your /home filesystem type autofs, which is ignored by mlocate by default. The speedup of automounting /home may not be more than a second or two, depending on your system, so this trick may not be worth it.

The same applies to remote filesystem mounts. If you want them to be mounted only upon access, you will need to use the noauto,x-systemd.automount parameters. In addition, you can use the x-systemd.device-timeout=# option to specify a timeout in case the network resource is not available.

If you have encrypted filesystems with keyfiles, you can also add the noauto parameter to the corresponding entries in /etc/crypttab. systemd will then not open the encrypted device on boot, but instead wait until it is actually accessed and then automatically open it with the specified keyfile before mounting it. This might save a few seconds on boot if you are using an encrypted RAID device for example, because systemd does not have to wait for the device to become available. For example:

/etc/crypttab
data /dev/md0 /root/key noauto

Filepath spaces

If any mountpoint contains spaces, use the escape character \ followed by the 3 digit octal code 040 to emulate them:

/etc/fstab
UUID=47FA-4071     /home/username/Camera\040Pictures   vfat  defaults,noatime       0  0
/dev/sda7          /media/100\040GB\040(Storage)       ext4  defaults,noatime,user  0  2

External devices

External devices that are to be mounted when present but ignored if absent may require the nofail option. This prevents errors being reported at boot.

/etc/fstab
/dev/sdg1        /media/backup    jfs    defaults,nofail    0  2

atime options

The use of noatime, nodiratime or relatime can impact drive performance.

  • The atime option updates the atime of the files every time they are accessed. This is more purposeful when Linux is used for servers; it does not have much value for desktop use. The drawback about the atime option is that even reading a file from the page cache (reading from memory instead of the drive) will still result in a write!
Using the noatime option fully disables writing file access times to the drive every time you read a file. This works well for almost all applications, except for a rare few like Mutt that needs such information. For mutt, you should only use the relatime option.
The nodiratime option disables the writing of file access times only for directories while other files still get access times written.
Note: noatime already includes nodiratime. You do not need to specify both.
  • relatime enables the writing of file access times only when the file is being modified (unlike noatime where the file access time will never be changed and will be older than the modification time). The best compromise might be the use this option since programs like Mutt will continue to work, but you will still have a performance boost as the files will not get access times updated unless they are modified. This option is used when the defaults keyword option or no options at all are specified in fstab for a given mount point.

tmpfs

tmpfs is a temporary filesystem that resides in memory and/or your swap partition(s), depending on how much you fill it up. Mounting directories as tmpfs can be an effective way of speeding up accesses to their files, or to ensure that their contents are automatically cleared upon reboot.

Some directories where tmpfs is commonly used are /tmp, /var/lock and /var/run. Do NOT use it on /var/tmp, because that folder is meant for temporary files that are preserved across reboots. Arch uses a tmpfs /run directory, with /var/run and /var/lock simply existing as symlinks for compatibility. It is also used for /tmp by the default systemd setup and does not require an entry in /etc/fstab unless a specific configuration is needed.

Note: When using systemd, temporary files in tmpfs directories can be recreated at boot by using tmpfiles.d.

By default, a tmpfs partition has its maximum size set to half your total RAM, but this can be customized. Note that the actual memory/swap consumption depends on how much you fill it up, as tmpfs partitions do not consume any memory until it is actually needed.

To explicitly set a maximum size, in this example to override the default /tmp mount, use the size mount option:

/etc/fstab
tmpfs   /tmp         tmpfs   nodev,nosuid,size=2G          0  0

Here is a more advanced example showing how to add tmpfs mounts for users. This is useful for websites, mysql tmp files, ~/.vim/, and more. It's important to try and get the ideal mount options for what you are trying to accomplish. The goal is to have as secure settings as possible to prevent abuse. Limiting the size, and specifying uid and gid + mode is very secure. For more information on this subject, follow the links listed in the #See also section.

/etc/fstab
tmpfs   /www/cache    tmpfs  rw,size=1G,nr_inodes=5k,noexec,nodev,nosuid,uid=648,gid=648,mode=1700   0  0

See the mount command man page for more information. One useful mount option in the man page is the default option. At least understand that.

Reboot for the changes to take effect. Note that although it may be tempting to simply run mount -a to make the changes effective immediately, this will make any files currently residing in these directories inaccessible (this is especially problematic for running programs with lockfiles, for example). However, if all of them are empty, it should be safe to run mount -a instead of rebooting (or mount them individually).

After applying changes, you may want to verify that they took effect by looking at /proc/mounts and using findmnt:

$ findmnt --target /tmp
TARGET SOURCE FSTYPE OPTIONS
/tmp   tmpfs  tmpfs  rw,nosuid,nodev,relatime

Usage

Generally, I/O intensive tasks and programs that run frequent read/write operations can benefit from using a tmpfs folder. Some applications can even receive a substantial gain by offloading some (or all) of their data onto the shared memory. For example, relocating the Firefox profile into RAM shows a significant improvement in performance.

Improving compile times

Compiling requires handling of many small files and involves many I/O operations; therefore it is a prime activity to benefit from moving its working directory to a #tmpfs.

For one session

The BUILDDIR value may be exported within a shell to temporarily set makepkg build directory to an existing tmpfs:

$ BUILDDIR=/tmp/makepkg makepkg
Permanently

Just uncomment the BUILDDIR line in /etc/makepkg.conf, see Makepkg#Improving compile times for details.

Writing to FAT32 as Normal User

Merge-arrows-2.pngThis article or section is a candidate for merging with USB Storage Devices#Mounting USB memory.Merge-arrows-2.png

Notes: The linked section assumes that the partition on USB storage uses FAT32 or NTFS filesystem, so we have two sections covering the same topic. Either merge everything here or in the linked section. (Discuss in Talk:Fstab#)

To write on a FAT32 partition, you must make a few changes to your /etc/fstab file.

/etc/fstab
/dev/sdxY    /mnt/some_folder  vfat   user,rw,umask=000              0  0

The user flag means that any user (even non-root) can mount and unmount the partition /dev/sdX. rw gives read-write access; umask option removes selected rights - for example umask=111 remove executable rights. The problem is that this entry removes executable rights from directories too, so we must correct it by dmask=000. See also Umask.

Without these options, all files will be executable. You can use the option showexec instead of the umask and dmask options, which shows all Windows executables (com, exe, bat) in executable colours.

For example, if your FAT32 partition is on /dev/sda9, and you wish to mount it to /mnt/fat32, then you would use:

/etc/fstab
/dev/sda9    /mnt/fat32        vfat   user,rw,umask=111,dmask=000    0  0

Remounting the root partition

If for some reason the root partition has been improperly mounted read only, remount the root partition with read-write access with the following command:

# mount -o remount,rw /

See also