1SYSTEMD-FSTAB-GENERATOR(8) systemd-fstab-generator SYSTEMD-FSTAB-GENERATOR(8)
2
3
4
6 systemd-fstab-generator - Unit generator for /etc/fstab
7
9 /usr/lib/systemd/system-generators/systemd-fstab-generator
10
12 systemd-fstab-generator is a generator that translates /etc/fstab (see
13 fstab(5) for details) into native systemd units early at boot and when
14 configuration of the system manager is reloaded. This will instantiate
15 mount and swap units as necessary.
16
17 The passno field is treated like a simple boolean, and the ordering
18 information is discarded. However, if the root file system is checked,
19 it is checked before all the other file systems.
20
21 See systemd.mount(5) and systemd.swap(5) for more information about
22 special /etc/fstab mount options this generator understands.
23
24 One special topic is handling of symbolic links. Historical init
25 implementations supported symlinks in /etc/fstab. Because mount units
26 will refuse mounts where the target is a symbolic link, this generator
27 will resolve any symlinks as far as possible when processing /etc/fstab
28 in order to enhance backwards compatibility. If a symlink target does
29 not exist at the time that this generator runs, it is assumed that the
30 symlink target is the final target of the mount.
31
32 systemd-fstab-generator implements systemd.generator(7).
33
35 systemd-fstab-generator understands the following kernel command line
36 parameters:
37
38 fstab=, rd.fstab=
39 Takes a boolean argument. Defaults to "yes". If "no", causes the
40 generator to ignore any mounts or swap devices configured in
41 /etc/fstab. rd.fstab= is honored only by the initial RAM disk
42 (initrd) while fstab= is honored by both the main system and the
43 initrd.
44
45 root=
46 Configures the operating system's root filesystem to mount when
47 running in the initrd. This accepts a device node path (usually
48 /dev/disk/by-uuid/... or /dev/disk/by-label/... or similar), or
49 the special values "gpt-auto" and "tmpfs".
50
51 Use "gpt-auto" to explicitly request automatic root file system
52 discovery via systemd-gpt-auto-generator(8).
53
54 Use "tmpfs" in order to mount a tmpfs(5) file system as root file
55 system of the OS. This is useful in combination with mount.usr=
56 (see below) in order to combine a volatile root file system with a
57 separate, immutable /usr/ file system. Also see systemd.volatile=
58 below.
59
60 rootfstype=
61 Takes the root filesystem type that will be passed to the mount
62 command. rootfstype= is honored by the initrd.
63
64 rootflags=
65 Takes the root filesystem mount options to use. rootflags= is
66 honored by the initrd.
67
68 Note that unlike most kernel command line options this setting does
69 not override settings made in configuration files (specifically:
70 the mount option string in /etc/fstab). See systemd-remount-
71 fs.service(8).
72
73 mount.usr=
74 Takes the /usr/ filesystem to be mounted by the initrd. If
75 mount.usrfstype= or mount.usrflags= is set, then mount.usr= will
76 default to the value set in root=.
77
78 Otherwise, this parameter defaults to the /usr/ entry found in
79 /etc/fstab on the root filesystem.
80
81 mount.usr= is honored by the initrd.
82
83 mount.usrfstype=
84 Takes the /usr/ filesystem type that will be passed to the mount
85 command. If mount.usr= or mount.usrflags= is set, then
86 mount.usrfstype= will default to the value set in rootfstype=.
87
88 Otherwise, this value will be read from the /usr/ entry in
89 /etc/fstab on the root filesystem.
90
91 mount.usrfstype= is honored by the initrd.
92
93 mount.usrflags=
94 Takes the /usr/ filesystem mount options to use. If mount.usr= or
95 mount.usrfstype= is set, then mount.usrflags= will default to the
96 value set in rootflags=.
97
98 Otherwise, this value will be read from the /usr/ entry in
99 /etc/fstab on the root filesystem.
100
101 mount.usrflags= is honored by the initrd.
102
103 systemd.volatile=
104 Controls whether the system shall boot up in volatile mode. Takes a
105 boolean argument or the special value state.
106
107 If false (the default), this generator makes no changes to the
108 mount tree and the system is booted up in normal mode.
109
110 If true the generator ensures systemd-volatile-root.service(8) is
111 run as part of the initial RAM disk ("initrd"). This service
112 changes the mount table before transitioning to the host system, so
113 that a volatile memory file system ("tmpfs") is used as root
114 directory, with only /usr/ mounted into it from the configured root
115 file system, in read-only mode. This way the system operates in
116 fully stateless mode, with all configuration and state reset at
117 boot and lost at shutdown, as /etc/ and /var/ will be served from
118 the (initially unpopulated) volatile memory file system.
119
120 If set to state the generator will leave the root directory mount
121 point unaltered, however will mount a "tmpfs" file system to /var/.
122 In this mode the normal system configuration (i.e. the contents of
123 "/etc/") is in effect (and may be modified during system runtime),
124 however the system state (i.e. the contents of "/var/") is reset at
125 boot and lost at shutdown.
126
127 If this setting is set to "overlay" the root file system is set up
128 as "overlayfs" mount combining the read-only root directory with a
129 writable "tmpfs", so that no modifications are made to disk, but
130 the file system may be modified nonetheless with all changes being
131 lost at reboot.
132
133 Note that in none of these modes the root directory, /etc/, /var/
134 or any other resources stored in the root file system are
135 physically removed. It's thus safe to boot a system that is
136 normally operated in non-volatile mode temporarily into volatile
137 mode, without losing data.
138
139 Note that with the exception of "overlay" mode, enabling this
140 setting will only work correctly on operating systems that can boot
141 up with only /usr/ mounted, and are able to automatically populate
142 /etc/, and also /var/ in case of "systemd.volatile=yes".
143
144 Also see root=tmpfs above, for a method to combine a "tmpfs" file
145 system with a regular /usr/ file system (as configured via
146 mount.usr=). The main distinction between systemd.volatile=yes, and
147 root=tmpfs in combination mount.usr= is that the former operates on
148 top of a regular root file system and temporarily obstructs the
149 files and directories above its /usr/ subdirectory, while the
150 latter does not hide any files, but simply mounts a unpopulated
151 tmpfs as root file system and combines it with a user picked /usr/
152 file system.
153
154 systemd.swap
155 Takes a boolean argument or enables the option if specified without
156 an argument. If disabled, causes the generator to ignore any swap
157 devices configured in /etc/fstab. Defaults to enabled.
158
160 systemd(1), fstab(5), systemd.mount(5), systemd.swap(5), systemd-
161 cryptsetup-generator(8), systemd-gpt-auto-generator(8), kernel-command-
162 line(7)
163
164
165
166systemd 249 SYSTEMD-FSTAB-GENERATOR(8)