mirror_zfs/man/man8/zfs-mount-generator.8.in
наб 0382362ce0 etc/systemd/zfs-mount-generator: rewrite in C
A plain rewrite of the shell version, and generates identical
units, save for replacing some empty lines with nothing, having fewer
meaningless spaces in After=s and different spacing in the lock scripts,
for a clean git diff -w

This is a gain of anywhere from 0m0.336s vs 0m0.022s (15.27x)
to 0m0.202s vs 0m0.006s (33.67x), depending on the hardware,
a.k.a. from "absolutely unusable" to "perfectly fine"

This also properly deals with canmount=noauto units across multiple
pools

See PR for detailed timings (of an early version) and diffs

Reviewed-by: Antonio Russo <aerusso@aerusso.net>
Reviewed-by: Richard Laager <rlaager@wiktel.com>
Reviewed-by: InsanePrawn <insane.prawny@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Issue #11915
Closes #11917
2021-06-08 14:50:38 -07:00

253 lines
6.9 KiB
Groff

.\"
.\" Copyright 2018 Antonio Russo <antonio.e.russo@gmail.com>
.\" Copyright 2019 Kjeld Schouten-Lebbing <kjeld@schouten-lebbing.nl>
.\" Copyright 2020 InsanePrawn <insane.prawny@gmail.com>
.\"
.\" Permission is hereby granted, free of charge, to any person obtaining
.\" a copy of this software and associated documentation files (the
.\" "Software"), to deal in the Software without restriction, including
.\" without limitation the rights to use, copy, modify, merge, publish,
.\" distribute, sublicense, and/or sell copies of the Software, and to
.\" permit persons to whom the Software is furnished to do so, subject to
.\" the following conditions:
.\"
.\" The above copyright notice and this permission notice shall be
.\" included in all copies or substantial portions of the Software.
.\"
.\" THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
.\" EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
.\" MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
.\" NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
.\" LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
.\" OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
.\" WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
.TH ZFS-MOUNT-GENERATOR 8 "Apr 19, 2021" OpenZFS
.SH "NAME"
zfs\-mount\-generator \- generates systemd mount units for ZFS
.SH SYNOPSIS
.B @systemdgeneratordir@/zfs\-mount\-generator
.sp
.SH DESCRIPTION
zfs\-mount\-generator implements the \fBGenerators Specification\fP
of
.BR systemd (1),
and is called during early boot to generate
.BR systemd.mount (5)
units for automatically mounted datasets. Mount ordering and dependencies
are created for all tracked pools (see below).
.SS ENCRYPTION KEYS
If the dataset is an encryption root, a service that loads the associated key (either from file or through a
.BR systemd\-ask\-password (1)
prompt) will be created. This service
. BR RequiresMountsFor
the path of the key (if file-based) and also copies the mount unit's
.BR After ,
.BR Before
and
.BR Requires .
All mount units of encrypted datasets add the key\-load service for their encryption root to their
.BR Wants
and
.BR After .
The service will not be
.BR Want ed
or
.BR Require d
by
.BR local-fs.target
directly, and so will only be started manually or as a dependency of a started mount unit.
.SS UNIT ORDERING AND DEPENDENCIES
mount unit's
.BR Before
\->
key\-load service (if any)
\->
mount unit
\->
mount unit's
.BR After
It is worth nothing that when a mount unit is activated, it activates all available mount units for parent paths to its mountpoint, i.e. activating the mount unit for /tmp/foo/1/2/3 automatically activates all available mount units for /tmp, /tmp/foo, /tmp/foo/1, and /tmp/foo/1/2. This is true for any combination of mount units from any sources, not just ZFS.
.SS CACHE FILE
Because ZFS pools may not be available very early in the boot process,
information on ZFS mountpoints must be stored separately. The output of the command
.PP
.RS 4
zfs list -H -o name,mountpoint,canmount,atime,relatime,devices,exec,readonly,setuid,nbmand,encroot,keylocation,org.openzfs.systemd:requires,org.openzfs.systemd:requires-mounts-for,org.openzfs.systemd:before,org.openzfs.systemd:after,org.openzfs.systemd:wanted-by,org.openzfs.systemd:required-by,org.openzfs.systemd:nofail,org.openzfs.systemd:ignore
.RE
.PP
for datasets that should be mounted by systemd, should be kept
separate from the pool at
.RI @sysconfdir@/zfs/zfs-list.cache/ POOLNAME .
.PP
The cache file, if writeable, will be kept synchronized with the pool
state by the
.I history_event-zfs-list-cacher.sh
ZEDLET.
.PP
.sp
.SS PROPERTIES
The behavior of the generator script can be influenced by the following dataset properties:
.sp
.TP 4
.BR canmount = on | off | noauto
If a dataset has
.BR mountpoint
set and
.BR canmount
is not
.BR off ,
a mount unit will be generated.
Additionally, if
.BR canmount
is
.BR on ,
.BR local-fs.target
will gain a dependency on the mount unit.
This behavior is equal to the
.BR auto
and
.BR noauto
legacy mount options, see
.BR systemd.mount (5).
Encryption roots always generate a key-load service, even for
.BR canmount=off .
.TP 4
.BR org.openzfs.systemd:requires\-mounts\-for = \fIpath\fR...
Space\-separated list of mountpoints to require to be mounted for this mount unit
.TP 4
.BR org.openzfs.systemd:before = \fIunit\fR...
The mount unit and associated key\-load service will be ordered before this space\-separated list of units.
.TP 4
.BR org.openzfs.systemd:after = \fIunit\fR...
The mount unit and associated key\-load service will be ordered after this space\-separated list of units.
.TP 4
.BR org.openzfs.systemd:wanted\-by = \fIunit\fR...
Space-separated list of units that will gain a
.BR Wants
dependency on this mount unit.
Setting this property implies
.BR noauto .
.TP 4
.BR org.openzfs.systemd:required\-by = \fIunit\fR...
Space-separated list of units that will gain a
.BR Requires
dependency on this mount unit.
Setting this property implies
.BR noauto .
.TP 4
.BR org.openzfs.systemd:nofail = unset | on | off
Toggles between a
.BR Wants
and
.BR Requires
type of dependency between the mount unit and
.BR local-fs.target ,
if
.BR noauto
isn't set or implied.
.BR on :
Mount will be
.BR WantedBy
local-fs.target
.BR off :
Mount will be
.BR Before
and
.BR RequiredBy
local-fs.target
.BR unset :
Mount will be
.BR Before
and
.BR WantedBy
local-fs.target
.TP 4
.BR org.openzfs.systemd:ignore = on | off
If set to
.BR on ,
do not generate a mount unit for this dataset.
See also
.BR systemd.mount (5)
.PP
.SH ENVIRONMENT
The
.BR $ZFS_DEBUG
environment variable, which can either be 0 (default),
1 (print summary accounting information at the end),
or at least 2 (print accounting information for each subprocess as it finishes).
If not present, /proc/cmdline is additionally checked for
.BR debug ,
in which case the debug level is set to 2.
.SH EXAMPLE
To begin, enable tracking for the pool:
.PP
.RS 4
touch
.RI @sysconfdir@/zfs/zfs-list.cache/ POOLNAME
.RE
.PP
Then, enable the tracking ZEDLET:
.PP
.RS 4
ln -s "@zfsexecdir@/zed.d/history_event-zfs-list-cacher.sh" "@sysconfdir@/zfs/zed.d"
systemctl enable zfs-zed.service
systemctl restart zfs-zed.service
.RE
.PP
Force the running of the ZEDLET by setting a monitored property, e.g.
.BR canmount ,
for at least one dataset in the pool:
.PP
.RS 4
zfs set canmount=on
.I DATASET
.RE
.PP
This forces an update to the stale cache file.
To test the generator output, run
.PP
.RS 4
@systemdgeneratordir@/zfs-mount-generator /tmp/zfs-mount-generator
.RE
.PP
This will generate units and dependencies in
.I /tmp/zfs-mount-generator
for you to inspect them. The second and third argument are ignored.
If you're satisfied with the generated units, instruct systemd to re-run all generators:
.PP
.RS 4
systemctl daemon-reload
.RE
.PP
.sp
.SH SEE ALSO
.BR zfs (5)
.BR zfs-events (5)
.BR zed (8)
.BR zpool (5)
.BR systemd (1)
.BR systemd.target (5)
.BR systemd.special (7)
.BR systemd.mount (7)