Strictly speaking you can run homed without userdb. But it doesn't
really make much sense: they go hand in hand and implement the same
concepts, just for different sets of users. Let's hence disable both
automatically by default if homed is requested.
(We don't do the reverse: opting into userdbd shouldn't mean that you
are OK with homed.)
And of course, users can always deviate from our defaults easily, and
turn off userbd again right-away if they don't like it, and things will
generally work.
This generator can be used by desktop environments to launch autostart
applications and services. The feature is an opt-in, triggered by
xdg-desktop-autostart.target being activated.
Also included is the new binary xdg-autostart-condition. This binary is
used as an ExecCondition to test the OnlyShowIn and NotShowIn XDG
desktop file keys. These need to be evaluated against the
XDG_CURRENT_DESKTOP environment variable which may not be known at
generation time.
Co-authored-by: Henri Chain <henri.chain@enioka.com>
In our regular gettys the actual shell commands live the the session
scope anyway (as long as logind is used). Hence, let's avoid
KillMode=process, it serves no purpose and is simply unsafe since it
disables systemd's own process lifecycle management.
We want to watch USB sticks being plugged in, and that requires
AF_NETLINK to work correctly and get the host's events. But if we live
in a network namespace AF_NETLINK is disconnected too and we'll not get
the host udev events.
Fixes: #15287
Limit size of various tmpfs mounts to 10% of RAM, except volatile root and /var
to 25%. Another exception is made for /dev (also /devs for PrivateDevices) and
/sys/fs/cgroup since no (or very few) regular files are expected to be used.
In addition, since directories, symbolic links, device specials and xattrs are
not counted towards the size= limit, number of inodes is also limited
correspondingly: 4MB size translates to 1k of inodes (assuming 4k each), 10% of
RAM (using 16GB of RAM as baseline) translates to 400k and 25% to 1M inodes.
Because nr_inodes option can't use ratios like size option, there's an
unfortunate side effect that with small memory systems the limit may be on the
too large side. Also, on an extremely small device with only 256MB of RAM, 10%
of RAM for /run may not be enough for re-exec of PID1 because 16MB of free
space is required.
"Login Service" doesn''t explain much, esp. considering that logind is actually is
for logins. I think "User Login Management" is better, but not that great either.
Suggestions welcome.
We unregister binfmt_misc twice during shutdown with this change:
1. A previous commit added support for doing that in the final shutdown
phase, i.e. when we do the aggressive umount loop. This is the robust
thing to do, in case the earlier ("clean") shutdown phase didn't work
for some reason.
2. This commit adds support for doing that when systemd-binfmt.service
is stopped. This is a good idea so that people can order mounts
before the service if they want to register binaries from such
mounts, as in that case we'll undo the registration on shutdown
again, before unmounting those mounts.
And all that, just because of that weird "F" flag the kernel introduced
that can pin files...
Fixes: #14981
This doesn't really matter, since in non-/usr-merged systems plymouth
needs to be in /bin and on merged ones it doesn't matter, but it is
still prettier to insert the right path, and avoid /bin on merged
systems, since it's just a compat symlink.
Replaces: #15351
This dependency is now generated automatically given we use
StateDirectory=. Moreover the combination of Wants= and After= was too
strong anway, as whether remount-fs is pulled in or not should not be up
to systemd-pstore.service, and in fact is part of the initial
transaction anyway.
sysinit.target is the target our early boot services are generally
pulled in from, make systemd-pstore.service not an exception of that.
Effectively this doesn't mean much, either way our unit is part of the
initial transaction.
Add `ProtectClock=yes` to systemd units. Since it implies certain
`DeviceAllow=` rules, make sure that the units have `DeviceAllow=` rules so
they are still able to access other devices. Exclude timesyncd and timedated.
This reverts commit 7e1ed1f3b2.
systemd-repart is not a user service that should be something people
enable/disable, instead it should just work if there's configuration for
it. It's like systemd-tmpfiles, systemd-sysusers, systemd-load-modules,
systemd-binfmt, systemd-systemd-sysctl which are NOPs if they have no
configuration, and thus don't hurt, but cannot be disabled since they
are too deep part of the OS.
This doesn't mean people couldn't disable the service if they really
want to, there's after all "systemctl mask" and build-time disabling,
but those are OS developer facing instead of admin facing, that's how it
should be.
Note that systemd-repart is in particular an initrd service, and so far
enable/disable state of those is not managed anyway via "systemctl
enable/disable" but more what dracut decides to package up and what not.
/home is posibly a remote file system. it makes sense to order homed
after it, so that we can properly enumerate users in it, but we probably
shouldn't pull it in ourselves, and leave that to users to configure
otherwise.
Fixes: #15102
It's lightweight and generally useful, so it should be enabled by default. But
users might want to disable it for whatever reason, and things should be fine
without it, so let's make it installable so it can be disabled if wanted.
Fixes#15175.
This essentially adds another layer of configurability:
build disable, this, presence of configuration. The default is
set to enabled, because the service does nothing w/o config.
Possible alternative to #14819.
For me, setting RemainAfterExit=yes would be OK, but if people think that it
might cause issues, then this could be a reasonable alternative that still
let's us skip the invocation of the separate binary.
This reverts the second part of 8125e8d38e.
The first part was reverted in 750e550eba.
The problem starts when s-v-s.s is pulled in by something that is then pulled
in by sysinit.target. Every time a unit is started, systemd recursively checks
all dependencies, and since sysinit.target is pull in by almost anything, we'll
start s-v-s.s over and over. In particular, plymouth-start.service currently
has Wants=s-v-s.s and After=s-v-s.s.
This minus has been there since the unit was added in
d42d27ead9. I think the idea was not cause things
to fail if the user instance doesn't work. But ignoring the return value
doesn't seem to be the right way to approach the problem. In particular, if
the program fails to run, we'll get a bogus fail state, see
https://bugzilla.redhat.com/show_bug.cgi?id=1727895#c1:
with the minus:
$ systemctl start user@1002
Job for user@1002.service failed because the service did not take the steps required by its unit configuration.
See "systemctl status user@1002.service" and "journalctl -xe" for details.
without the minus:
$ systemctl start user@1002
Job for user@1002.service failed because the control process exited with error code.
See "systemctl status user@1002.service" and "journalctl -xe" for details.
This patch modifies the RequireMountsFor setting in systemd-nspawn@.service to wait for the machine instance directory to be mounted, not just /var/lib/machines.
Closes#14931
machined needs access to the host mount namespace to propagate bind
mounts created with the "machinectl bind" command. However, the
"ProtectKernelLogs" directive relies on mount namespaces to make the
kernel ring buffer inaccessible. This commit removes the
"ProtectKernelLogs=yes" directive from machined service file introduced
in 6168ae5.
Closes#14559.
Kernel 4.1 separated the tracing system from the debugfs,
actual documentation already points to a different path
that needs this new mount to exist.
the old sysfs path will still be an automount in the debugfs,
created by the kernel (for now).
Signed-off-by: Norbert Lange <nolange79@gmail.com>
See c80a9a33d0, target units can't fail.
I guess we need to figure out some replacement functionality, but at least
let's avoid the warning from systemd for now.
If we have exit on idle, then operations such as "journalctl
--namespace=foo --rotate" should work even if the journal daemon is
currently not running.
(Note that we don't do activation by varlink for the main instance of
journald, I am not sure the deadlocks it might introduce are worth it)
This makes things a bit simpler and the build a bit faster, because we don't
have to rewrite files to do the trivial substitution. @rootbindir@ is always in
our internal $PATH that we use for non-absolute paths, so there should be no
functional change.
Let's use uppercase wording in the description string, like we usually
do.
Let's allow using this service in early boot.
If it's pulled into the initial transaction it's better to finish
loading this before sysinit.target.
Don't bother with this in containers that lack CAP_SYS_MODULE
Devices referred to by `DeviceAllow=` sandboxing are resolved into their
corresponding major numbers when the unit is loaded by looking at
`/proc/devices`. If a reference is made to a device which is not yet
available, the `DeviceAllow` is ignored and the unit's processes cannot
access that device.
In both logind and nspawn, we have `DeviceAllow=` lines, and `modprobe`
in `ExecStartPre=` to load some kernel modules. Those kernel modules
cause device nodes to become available when they are loaded: the device
nodes may not exist when the unit itself is loaded. This means that the
unit's processes will not be able to access the device since the
`DeviceAllow=` will have been resolved earlier and denied it.
One way to fix this would be to re-evaluate the available devices and
re-apply the policy to the cgroup, but this cannot work atomically on
cgroupsv1. So we fall back to a second approach: instead of running
`modprobe` via `ExecStartPre`, we move this out to a separate unit and
order it before the units which want the module.
Closes#14322.
Fixes: #13943.
This reverts commit 07125d24ee.
In contrast to what is claimed in #13396 dbus-broker apparently does
care for the service file to be around, and otherwise will claim
"Service Not Activatable" in the time between systemd starting up the
broker and connecting to it, which the stub service file is supposed to
make go away.
Reverting this makes the integration test suite pass again on host with
dbus-broker (i.e. current Fedora desktop).
Tested with dbus-broker-21-6.fc31.x86_64.
This reverts commit 362c378291.
This commit introduced an ordering loop: remote-cryptsetup.target was both
before and after remote-fs-pre.target. It also globally ordered all cryptsetup
volumes before all mounts. Such global ordering is problematic if people have
stacked storage. Let's look for a different solution.
See https://github.com/systemd/systemd/pull/14378#discussion_r359460109.
Otherwise, systemd-udev-trigger|settle.service that ran in the initrd may
ramain active, and never re-run again from the system root.
This is observed by forexample examining ESP with udevadm info, which in the
initrd has all the ID_* variables, and none of them in fully booted system.
This option is an indication for PID1 that the entry in crypttab is handled by
initrd only and therefore it shouldn't interfer during the usual start-up and
shutdown process.
It should be primarily used with the encrypted device containing the root FS as
we want to keep it (and thus its encrypted device) until the very end of the
shutdown process, i.e. when initrd takes over.
This option is the counterpart of "x-initrd.mount" used in fstab.
Note that the slice containing the cryptsetup services also needs to drop the
usual shutdown dependencies as it's required by the cryptsetup services.
Fixes: #14224
Apparently some firmwares don't allow us to write this token, and refuse
it with EINVAL. We should normally consider that a fatal error, but not
really in the case of "bootctl random-seed" when called from the
systemd-boot-system-token.service since it's called as "best effort"
service after boot on various systems, and hence we shouldn't fail
loudly.
Similar, when we cannot find the ESP don't fail either, since there are
systems (arch install ISOs) that carry a boot loader capable of the
random seed logic but don't mount it after boot.
Fixes: #13603
We set ProtectKernelLogs=yes on all long running services except for
udevd, since it accesses /dev/kmsg, and journald, since it calls syslog
and accesses /dev/kmsg.
As discussed on systemd-devel [1], in Fedora we get lots of abrt reports
about the watchdog firing [2], but 100% of them seem to be caused by resource
starvation in the machine, and never actual deadlocks in the services being
monitored. Killing the services not only does not improve anything, but it
makes the resource starvation worse, because the service needs cycles to restart,
and coredump processing is also fairly expensive. This adds a configuration option
to allow the value to be changed. If the setting is not set, there is no change.
My plan is to set it to some ridiculusly high value, maybe 1h, to catch cases
where a service is actually hanging.
[1] https://lists.freedesktop.org/archives/systemd-devel/2019-October/043618.html
[2] https://bugzilla.redhat.com/show_bug.cgi?id=1300212
The code supports SIGTERM and SIGINT to termiante the process. It would
be possible to reporpose one of those signals for the restart operation,
but I think it's better to use a completely different signal to avoid
misunderstandings.
See https://bugzilla.redhat.com/show_bug.cgi?id=1731772:
when autofs4 is disabled in the kernel,
proc-sys-fs-binfmt_misc.automount is not started, so the binfmt_misc module is
never loaded. If we added a dependency on proc-sys-fs-binfmt_misc.mount
to systemd-binfmt.service, things would work even if autofs4 was disabled, but
we would unconditionally pull in the module and mount, which we don't want to do.
(Right now we ony load the module if some binfmt is configured.)
But let's make it easier to handle this case by doing two changes:
1. order systemd-binfmt.service after the .mount unit (so that the .service
can count on the mount if both units are pulled in, even if .automount
is skipped)
2. add [Install] section to the service unit. This way the user can do
'systemctl enable proc-sys-fs-binfmt_misc.mount' to get the appropriate behaviour.
This fixes the following problem:
> At the very end of the boot, just after the first user logs in
> (usually using sddm / X) I get the following messages in my logs:
> Nov 18 07:02:33 samd dbus-daemon[2879]: [session uid=1000 pid=2877] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
> Nov 18 07:02:33 samd dbus-daemon[2879]: [session uid=1000 pid=2877] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
These messages are caused by the "stub" service files that systemd
installs. It installed them because early versions of systemd activation
required them to exist.
Since dbus 1.11.0, a dbus-daemon that is run with --systemd-activation
automatically assumes that o.fd.systemd1 is an activatable
service. As a result, with a new enough dbus version,
/usr/share/dbus-1/services/org.freedesktop.systemd1.service and
/usr/share/dbus-1/system-services/org.freedesktop.systemd1.service should
become unnecessary, and they can be removed.
dbus 1.11.0 was released 2015-12-02.
Bug-Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=914015
If logging disappears issues are hard to debug, hence let's give
journald a slight edge over other services when the OOM killer hits.
Here are the special adjustments we now make:
systemd-coredump@.service.in OOMScoreAdjust=500
systemd-journald.service.in OOMScoreAdjust=-250
systemd-udevd.service.in OOMScoreAdjust=-1000
(i.e. the coredump processing is made more likely to be killed on OOM,
and udevd and journald are less likely to be killed)
Follow-up for 26ded55709.
The commit says,
> Note that with this change sysinit.target (and thus early boot) is NOT
systematically delayed until the entropy pool is initialized,
But the dependency was not dropped.
This was found by David Seifert (@SoapGentoo).
This makes two major changes to the way systemd-random-seed operates:
1. We now optionally credit entropy if this is configured (via an env
var). Previously we never would do that, with this change we still don't
by default, but it's possible to enable this if people acknowledge that
they shouldn't replicate an image with a contained random seed to
multiple systems. Note that in this patch crediting entropy is a boolean
thing (unlike in previous attempts such as #1062), where only a relative
amount of bits was credited. The simpler scheme implemented here should
be OK though as the random seeds saved to disk are now written only with
data from the kernel's entropy pool retrieved after the pool is fully
initialized. Specifically:
2. This makes systemd-random-seed.service a synchronization point for
kernel entropy pool initialization. It was already used like this, for
example by systemd-cryptsetup-generator's /dev/urandom passphrase
handling, with this change it explicitly operates like that (at least
systems which provide getrandom(), where we can support this). This
means services that rely on an initialized random pool should now place
After=systemd-random-seed.service and everything should be fine. Note
that with this change sysinit.target (and thus early boot) is NOT
systematically delayed until the entropy pool is initialized, i.e.
regular services need to add explicit ordering deps on this service if
they require an initialized random pool.
Fixes: #4271
Replaces: #10621#4513
This reverts commit 971a7a1526.
These unit names are typically different on distributions, let's not
hardcode those. Stuff like this should probably live in the distro
RPM/.deb, but not upstream, where we should be distro agnostic and
agnostic to other higher level packages like this.
Users might end up with more than one of those service enabled, through
admin mistake, or broken installation scriptlets, or whatever. On my machine,
I had both chronyd and timesyncd happilly running at the same time. If
more than one is enabled, it's better to have just one running. Adding
Conflicts will make the issue more visible in logs too.
This patch introduces the systemd pstore service which will archive the
contents of the Linux persistent storage filesystem, pstore, to other storage,
thus preserving the existing information contained in the pstore, and clearing
pstore storage for future error events.
Linux provides a persistent storage file system, pstore[1], that can store
error records when the kernel dies (or reboots or powers-off). These records in
turn can be referenced to debug kernel problems (currently the kernel stuffs
the tail of the dmesg, which also contains a stack backtrace, into pstore).
The pstore file system supports a variety of backends that map onto persistent
storage, such as the ACPI ERST[2, Section 18.5 Error Serialization] and UEFI
variables[3 Appendix N Common Platform Error Record]. The pstore backends
typically offer a relatively small amount of persistent storage, e.g. 64KiB,
which can quickly fill up and thus prevent subsequent kernel crashes from
recording errors. Thus there is a need to monitor and extract the pstore
contents so that future kernel problems can also record information in the
pstore.
The pstore service is independent of the kdump service. In cloud environments
specifically, host and guest filesystems are on remote filesystems (eg. iSCSI
or NFS), thus kdump relies [implicitly and/or explicitly] upon proper operation
of networking software *and* hardware *and* infrastructure. Thus it may not be
possible to capture a kernel coredump to a file since writes over the network
may not be possible.
The pstore backend, on the other hand, is completely local and provides a path
to store error records which will survive a reboot and aid in post-mortem
debugging.
Usage Notes:
This tool moves files from /sys/fs/pstore into /var/lib/systemd/pstore.
To enable kernel recording of error records into pstore, one must either pass
crash_kexec_post_notifiers[4] to the kernel command line or enable via 'echo Y
> /sys/module/kernel/parameters/crash_kexec_post_notifiers'. This option
invokes the recording of errors into pstore *before* an attempt to kexec/kdump
on a kernel crash.
Optionally, to record reboots and shutdowns in the pstore, one can either pass
the printk.always_kmsg_dump[4] to the kernel command line or enable via 'echo Y >
/sys/module/printk/parameters/always_kmsg_dump'. This option enables code on the
shutdown path to record information via pstore.
This pstore service is a oneshot service. When run, the service invokes
systemd-pstore which is a tool that performs the following:
- reads the pstore.conf configuration file
- collects the lists of files in the pstore (eg. /sys/fs/pstore)
- for certain file types (eg. dmesg) a handler is invoked
- for all other files, the file is moved from pstore
- In the case of dmesg handler, final processing occurs as such:
- files processed in reverse lexigraphical order to faciliate
reconstruction of original dmesg
- the filename is examined to determine which dmesg it is a part
- the file is appended to the reconstructed dmesg
For example, the following pstore contents:
root@vm356:~# ls -al /sys/fs/pstore
total 0
drwxr-x--- 2 root root 0 May 9 09:50 .
drwxr-xr-x 7 root root 0 May 9 09:50 ..
-r--r--r-- 1 root root 1610 May 9 09:49 dmesg-efi-155741337601001
-r--r--r-- 1 root root 1778 May 9 09:49 dmesg-efi-155741337602001
-r--r--r-- 1 root root 1726 May 9 09:49 dmesg-efi-155741337603001
-r--r--r-- 1 root root 1746 May 9 09:49 dmesg-efi-155741337604001
-r--r--r-- 1 root root 1686 May 9 09:49 dmesg-efi-155741337605001
-r--r--r-- 1 root root 1690 May 9 09:49 dmesg-efi-155741337606001
-r--r--r-- 1 root root 1775 May 9 09:49 dmesg-efi-155741337607001
-r--r--r-- 1 root root 1811 May 9 09:49 dmesg-efi-155741337608001
-r--r--r-- 1 root root 1817 May 9 09:49 dmesg-efi-155741337609001
-r--r--r-- 1 root root 1795 May 9 09:49 dmesg-efi-155741337710001
-r--r--r-- 1 root root 1770 May 9 09:49 dmesg-efi-155741337711001
-r--r--r-- 1 root root 1796 May 9 09:49 dmesg-efi-155741337712001
-r--r--r-- 1 root root 1787 May 9 09:49 dmesg-efi-155741337713001
-r--r--r-- 1 root root 1808 May 9 09:49 dmesg-efi-155741337714001
-r--r--r-- 1 root root 1754 May 9 09:49 dmesg-efi-155741337715001
results in the following:
root@vm356:~# ls -al /var/lib/systemd/pstore/155741337/
total 92
drwxr-xr-x 2 root root 4096 May 9 09:50 .
drwxr-xr-x 4 root root 40 May 9 09:50 ..
-rw-r--r-- 1 root root 1610 May 9 09:50 dmesg-efi-155741337601001
-rw-r--r-- 1 root root 1778 May 9 09:50 dmesg-efi-155741337602001
-rw-r--r-- 1 root root 1726 May 9 09:50 dmesg-efi-155741337603001
-rw-r--r-- 1 root root 1746 May 9 09:50 dmesg-efi-155741337604001
-rw-r--r-- 1 root root 1686 May 9 09:50 dmesg-efi-155741337605001
-rw-r--r-- 1 root root 1690 May 9 09:50 dmesg-efi-155741337606001
-rw-r--r-- 1 root root 1775 May 9 09:50 dmesg-efi-155741337607001
-rw-r--r-- 1 root root 1811 May 9 09:50 dmesg-efi-155741337608001
-rw-r--r-- 1 root root 1817 May 9 09:50 dmesg-efi-155741337609001
-rw-r--r-- 1 root root 1795 May 9 09:50 dmesg-efi-155741337710001
-rw-r--r-- 1 root root 1770 May 9 09:50 dmesg-efi-155741337711001
-rw-r--r-- 1 root root 1796 May 9 09:50 dmesg-efi-155741337712001
-rw-r--r-- 1 root root 1787 May 9 09:50 dmesg-efi-155741337713001
-rw-r--r-- 1 root root 1808 May 9 09:50 dmesg-efi-155741337714001
-rw-r--r-- 1 root root 1754 May 9 09:50 dmesg-efi-155741337715001
-rw-r--r-- 1 root root 26754 May 9 09:50 dmesg.txt
where dmesg.txt is reconstructed from the group of related
dmesg-efi-155741337* files.
Configuration file:
The pstore.conf configuration file has four settings, described below.
- Storage : one of "none", "external", or "journal". With "none", this
tool leaves the contents of pstore untouched. With "external", the
contents of the pstore are moved into the /var/lib/systemd/pstore,
as well as logged into the journal. With "journal", the contents of
the pstore are recorded only in the systemd journal. The default is
"external".
- Unlink : is a boolean. When "true", the default, then files in the
pstore are removed once processed. When "false", processing of the
pstore occurs normally, but the pstore files remain.
References:
[1] "Persistent storage for a kernel's dying breath",
March 23, 2011.
https://lwn.net/Articles/434821/
[2] "Advanced Configuration and Power Interface Specification",
version 6.2, May 2017.
https://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf
[3] "Unified Extensible Firmware Interface Specification",
version 2.8, March 2019.
https://uefi.org/sites/default/files/resources/UEFI_Spec_2_8_final.pdf
[4] "The kernel’s command-line parameters",
https://static.lwn.net/kerneldoc/admin-guide/kernel-parameters.html
We use that on all other services, and hence should here too. Otherwise
the service will be killed with SIGSYS when doing something not
whitelisted, which is a bit crass.
While the need for access to character devices can be tricky to determine for
the general case, it's obvious that most of our services have no need to access
block devices. For logind and timedated this can be tightened further.
We had all kinds of indentation: 2 sp, 3 sp, 4 sp, 8 sp, and mixed.
4 sp was the most common, in particular the majority of scripts under test/
used that. Let's standarize on 4 sp, because many commandlines are long and
there's a lot of nesting, and with 8sp indentation less stuff fits. 4 sp
also seems to be the default indentation, so this will make it less likely
that people will mess up if they don't load the editor config. (I think people
often use vi, and vi has no support to load project-wide configuration
automatically. We distribute a .vimrc file, but it is not loaded by default,
and even the instructions in it seem to discourage its use for security
reasons.)
Also remove the few vim config lines that were left. We should either have them
on all files, or none.
Also remove some strange stuff like '#!/bin/env bash', yikes.
time-sync.target is supposed to indicate system clock is synchronized
with a remote clock, but as used through 241 it only provided a system
clock that was updated based on a locally-maintained timestamp. Systems
that are powered off for extended periods would not come up with
accurate time.
Retain the existing behavior using a new time-set.target leaving
time-sync.target for cases where accuracy is required.
Closes#8861
Let's be safe, rather than sorry. This way DynamicUser=yes services can
neither take benefit of, nor create SUID/SGID binaries.
Given that DynamicUser= is a recent addition only we should be able to
get away with turning this on, even though this is strictly speaking a
binary compatibility breakage.
This patch was initially prompted by a report on a Fedora update [1], that the
upgrade causes systemd-resolved.service and systemd-networkd.service to be
re-enabled. We generally want to preserve the enablement of all services during
upgrades, so a reset like this is not expected.
Both services declare two symlinks in their [Install] sections, for their dbus
names and for multi-user.target.wants/. It turns out that both services were
only partially enabled, because their dbus unit symlinks
/etc/systemd/system/dbus-org.freedesktop.{resolve1,network1}.service were
created, by the symlinks in /etc/systemd/system/multi-user.target.wants/ were
not. This means that the units could be activated by dbus, but not in usual
fashion using systemctl start. Our tools make it rather hard to figure out when
something like this happens, and it is definitely an area for improvement on its
own. The symlink in .wants/ was filtered out by during packaging, but the dbus
symlink was left in (I assume by mistake).
Let's simplify things by not creating the symlinks statically during 'ninja
install'. This means that the units shipped by systemd have to be enabled in
the usual fashion, which in turns means that [Install] section and presets
become the "single source of truth" and we don't have two sets of conflicting
configuration.
Let's consider a few cases:
- developer: a developer installs systemd from git on a running system, and they
don't want the installation to reset enablement of anything. So this change is
either positive for them, or has no effect (if they have everything at
defaults).
- package creation: we want to create symlinks using 'preset-all' and 'preset'
on upgraded packages, we don't want to have any static symlinks. This change
will remove the need to filter out symlinks in packaging and of course fix
the original report.
- installation of systemd from scratch: this change means that without
'preset-all' the system will not be functional. This case could be affected
negatively by this change, but I think it's enough of a corner case to accept
this. In practice I expect people to build a package, not installl directly
into the file system, so this might not even matter in practice.
Creating those symlinks was probably the right thing in the beginning, but
nowadays the preset system is very well established and people expect it to
be honoured. Ignoring the presets and doing static configuration is not welcome
anymore.
Note: during package installation, either 'preset-all' or 'preset getty@.service
machines.target remote-cryptsetup.target remote-fs.target
systemd-networkd.service systemd-resolved.service
systemd-networkd-wait-online.service systemd-timesyncd.service' should be called.
[1] https://bodhi.fedoraproject.org/updates/FEDORA-2019-616045ca76
A couple of API VFS we mount via .mount units. Let's set the three flags
for those too, just in case.
This is just paranoia, nothing else, but shouldn't hurt.
This service uses PAM anyway, hence let pam_keyring set things up for
us. Moreover, this way we ensure that the invocation ID is not set for
this service as key, and thus can't confuse the user service's
invocation ID.
Fixes: #11649
`systemd-journal-catalog-update.service` writes to `/var`. However, it's
not explicitly ordered wrt `systemd-tmpfiles-setup.service`, which means
that it may run before or after.
This is an issue for Fedora CoreOS, which uses Ignition. We want to be
able to prepare `/var` on first boot from the initrd, where the SELinux
policy is not loaded yet. This means that the hierarchy under `/var` is
not correctly labeled. We add a `Z /var - - -` tmpfiles entry so that it
gets relabeled once `/var` gets mounted post-switchroot.
So any service that tries to access `/var` before `systemd-tmpfiles`
relabels it is likely to hit `EACCES`.
Fix this by simply ordering `systemd-journal-catalog-update.service`
after `systemd-tmpfiles-setup.service`. This is also clearer since the
tmpfiles entries are the canonical source of how `/var` should be
populated.
For more context on this, see:
https://github.com/coreos/ignition/issues/635#issuecomment-446620297
ProtectHostname= turns off hostname change propagation from host to
service. This means for services that care about the hostname and need
to be able to notice changes to it it's not suitable (though it is
useful for most other cases still).
Let's turn it off hence for journald (which logs the current hostname)
for networkd (which optionally sends the current hostname to dhcp
servers) and resolved (which announces the current hostname via
llmnr/mdns).
This behaves similar to the "boot into firmware" logic, and also allows
either direct EFI operation (which sd-boot supports and others might
support eventually too) or override through env var.
This was an overzealous setting from commit 99894b867f. Without this,
`hostnamectl set-hostname` fails with
Could not set property: Access denied
as `sethostname()` fails with `EPERM`.
Linux can be run on a device meant to act as a USB peripheral. In order
for a machine to act as such a USB device it has to be equipped with
a UDC - USB Device Controller.
This patch adds a target reached when UDC becomes available. It can be used
for activating e.g. a service unit which composes a USB gadget with
configfs and activates it.
A follow-up for commit a8cb1dc3e0.
Commit a8cb1dc3e0 made sure that initrd-cleanup.service won't be stopped
when initrd-switch-root.target is isolated.
However even with this change, it might happen that initrd-cleanup.service
survives the switch to rootfs (since it has no ordering constraints against
initrd-switch-root.target) and is stopped right after when default.target is
isolated. This led to initrd-cleanup.service entering in failed state as it
happens when oneshot services are stopped.
This patch along with a8cb1dc3e0 should fix issue #4343.
Fixes: #4343
Currently, tmpfiles runs in two separate services at boot. /dev is
populated by systemd-tmpfiles-setup-dev.service and everything else by
systemd-tmpfiles-setup.service. The former was so far conditionalized by
CAP_SYS_MODULES. The reasoning was that the primary purpose of
populating /dev was to create device nodes based on the static device
node info exported in kernel modules through MODALIAS. And without the
privs to load kernel modules doing so is unnecessary. That thinking is
incomplete however, as there might be reason to create stuff in /dev
outside of the static modalias usecase. Thus, let's drop the
conditionalization to ensure that tmpfiles.d rules are always executed
at least once under all conditions.
Fixes: #11544
Instead of enabling it unconditionally and then using ConditionPathExists=/etc/fstab,
and possibly masking this condition if it should be enabled for auto gpt stuff,
just pull it in explicitly when required.
We already *install* those as real files since de78fa9ba0.
Meson will start to copy symlinks as-is, so we would get dangling symlinks in
/usr/lib/systemd/user/.
I considered the layout in our sources to match the layout in the installation
filesystem (i.e. creating units/system/ and moving all files from units/ to
units/system/), but that seems overkill. By using normal files for both we get
some duplication, but those files change rarely, so it's not a big downside in
practice.
Fixes#9906.
Let's simplify things and drop the logic that /var/lib/machines is setup
as auto-growing btrfs loopback file /var/lib/machines.raw.
THis was done in order to make quota available for machine management,
but quite frankly never really worked properly, as we couldn't grow the
file system in sync with its use properly. Moreover philosophically it's
problematic overriding the admin's choice of file system like this.
Let's hence drop this, and simplify things. Deleting code is a good
feeling.
Now that regular file systems provide project quota we could probably
add per-machine quota support based on that, hence the btrfs quota
argument is not that interesting anymore (though btrfs quota is a bit
more powerful as it allows recursive quota, i.e. that the machine pool
gets an overall quota in addition to per-machine quota).
Otherwise we might install the socket unit early, but the service
backing it late, and then end up in strange loops when we enter rescue
mode, because we saw an event on /dev/rfkill but really can't dispatch
it nor flush it.
Fixes: #9171
now that logind doesn't mount $XDG_RUNTIME_DIR anymore we can lock down
the service using fs namespacing (as we don't need the mount to
propagate to the host namespace anymore).
Previously, setting this option by default was problematic due to
SELinux (as this would also prohibit the transition from PID1's label to
the service's label). However, this restriction has since been lifted,
hence let's start making use of this universally in our services.
On SELinux system this change should be synchronized with a policy
update that ensures that NNP-ful transitions from init_t to service
labels is permitted.
An while we are at it: sort the settings in the unit files this touches.
This might increase the size of the change in this case, but hopefully
should result in stabler patches later on.
Fixes: #1219
I found zero references to busnames.target, using git grep "busnames".
(And we do not install using a wildcard units/*.*. There is no
busnames.target installed on my Fedora 28 system).
THis dep existed since the unit was introduced, but I cannot see what
good it would do. Hence in the interest of simplifying things, let's
drop it. If breakages appear later we can certainly revert this again.
Fixes: #10469
This is might be useful in some cases, but it's primarily an example for
a boot check service that can be plugged before boot-complete.target.
It's disabled by default.
All it does is check whether the failed unit count is zero
This is the counterpiece to the boot counting implemented in
systemd-boot: if a boot is detected as successful we mark drop the
counter again from the booted snippet or kernel image.
C.f. 287419c119: 'systemctl exit 42' can be
used to set an exit value and pulls in exit.target, which pulls in systemd-exit.service,
which calls org.fdo.Manager.Exit, which calls method_exit(), which sets the objective
to MANAGER_EXIT. Allow the same to happen through SuccessAction=exit.
v2: update for 'exit' and 'exit-force'
Explicit systemctl calls remain in systemd-halt.service and the system
systemd-exit.service. To convert systemd-halt, we'd need to add
SuccessAction=halt-force. Halting doesn't make much sense, so let's just
leave that is. systemd-exit.service will be converted in the next commit.
This updates the unit files of all our serviecs that deal with journal
stuff to use a higher RLIMIT_NOFILE soft limit by default. The new value
is the same as used for the new HIGH_RLIMIT_NOFILE we just added.
With this we ensure all code that access the journal has higher
RLIMIT_NOFILE. The code that runs as daemon via the unit files, the code
that is run from the user's command line via C code internal to the
relevant tools. In some cases this means we'll redundantly bump the
limits as there are tools run both from the command line and as service.
So far we always used "yes" instead of "true" in all our unit files,
except for one outlier. Let's do this here too. No change in behaviour
whatsoever, except that it looks prettier ;-)
I think this is a slightly cleaner approach than parsing the
configuration file at multiple places, as this way there's only a single
reload cycle for logind.conf, and that's systemd-logind.service's
runtime.
This means that logind and dbus become a requirement of
user-runtime-dir, but given that XDG_RUNTIME_DIR is not set anyway
without logind and dbus around this isn't really any limitation.
This also simplifies linking a bit as this means user-runtime-dir
doesn't have to link against any code of logind itself.
Let's not use the word "wrapper", as it's not clear what that is, and in
some way any unit file is a "wrapper"... let's simply say that it's
about the runtime directory.
If for any reason local-fs.target fails at startup while a password is
requested by systemd-cryptsetup@.service, we end up with the emergency shell
competing with systemd-ask-password-console.service for the console.
This patch makes sure that:
- systemd-ask-password-console.service is stopped before entering in emergency
mode so it won't make any access to the console while the emergency shell is
running.
- systemd-ask-password-console.path is also stopped so any attempts to restart
systemd-cryptsetup in the emergency shell won't restart
systemd-ask-password-console.service and kill the emergency shell.
- systemd-ask-password-wall.path is stopped so
systemd-ask-password-wall.service won't be started as this service pulls
the default dependencies in.
Fixes: #10131
This reverts commit d4e9e574ea.
(systemd.conf.m4 part was already reverted in 5b5d82615011b9827466b7cd5756da35627a1608.)
Together those reverts should "fix" #10025 and #10011. ("fix" is in quotes
because this doesn't really fix the underlying issue, which is combining
DynamicUser= with strict container sandbox, but it avoids the problem by not
using that feature in our default installation.)
Dynamic users don't work well if the service requires matching configuration in
other places, for example dbus policy. This is true for those three services.
In effect, distros create the user statically [1, 2]. Dynamic users make more
sense for "add-on" services where not creating the user, or more precisely,
creating the user lazily, can save resources. For "basic" services, if we are
going to create the user on package installation anyway, setting DynamicUser=
just creates unneeded confusion. The only case where it is actually used is
when somebody forgets to do system configuration. But it's better to have the
service fail cleanly in this case too. If we want to turn on some side-effect
of DynamicUser=yes for those services, we should just do that directly through
fine-grained options. By not using DynamicUser= we also avoid the need to
restart dbus.
[1] bd9bf30727
[2] 48ac1cebde/f/systemd.spec (_473)
(Fedora does not create systemd-timesync user.)
systemd-tmpfiles-setup.service needs to be ordered after
systemd-journald.service, so entries in /run/log/journal are already
created when systemd-tmpfiles tries to adjust its permissions.
This is specially problematic for setups using a volatile journal where
the initrd does not ship a machine-id (i.e. OSTree-based systems), where
logs from the initrd will be inaccessible for users in the
systemd-journal group. It also has a side effect of `journalctl --user`
failing with "No journal files were opened due to insufficient
permissions".
Fixes#10128.
We would create a useless empty directory under build/.
It seems we were lucky and all symlinks were installed into directories
which were alredy created because we installed something into the same
location earlier.
While at it, also add '-v' to 'mkdir -p'. This will print the names of
directories as they are created (just once), making it easier to see all of
what the install script is doing.
This reverts commit 48d3e88c18.
I kept the follow-symlink=false → follow-symlink=true change instact, since
we're likely to have existing installations with a symlink now.
Followup to commit 13cf422e04 ("user@.service: don't kill user manager at runlevel switch")
I think there's a general rule that units with `StopWhenUnneeded=yes` need
`IgnoreOnIsolate=yes`... But it doesn't apply to `suspend.target` and friends.
`printer.target` and friends break on isolate even if we apply the rule[1].
That just leaves `graphical-session.target`, which is a user service.
"isolate" is *mostly* a weird attempt to emulate runlevels, so I decided
not to worry about it for user services.
[1] https://github.com/systemd/systemd/issues/6505#issuecomment-320644819
Loggin in as root user and then switching the runlevel results in a
stop of the user manager, even though the user ist still logged in.
That leaves a broken user session.
Adding "IgnoreOnIsolate=true" to user@.service fixes this.
This service won't use much resources, but it's certainly nicer to see
it attached th the user's slice along with user@.service, so that
everything we run for a specific user is properly bound into one unit.
We use systemd-user-sessions.service as barrier when to allow login
sessions. With this patch user@.service is ordered after that too, so
that any login related code (which user-runtime-dir@.service is) is
guaranteed to run after the barrier, and never before.
This is an additional synchronization point normally not needed. Hence,
let's make it passive, i.e. pull it in from the unit which wants to be
ordered before the update service rather than by the update service
itself.
systemd offline-updates allows dropping multiple system update units
to be added to system-update.target.wants.
As documented in systemd.offline-updates(7) only 1 of these units
should actually be active (based on the /system-update symlink) and
when that unit is done it should reboot the system.
In some cases it is desirable to run a unit whenever booting in
offline-updates mode indepedent of which update unit is going to
handle the update. One example of this is integration with bootloader
code which checks if the previous boot was succesful.
Since the active unit will reboot the system when it is done, there
is no guarantee that adding such a unit to system-update.target.wants
will get it executed always.
This commit adds a system-update-pre.target which can be used for
units which should always run when booting in offline-updates mode.
This is generally the safer approach, and is what container managers
(including nspawn) do, hence let's move to this too for our own
services. This is particularly useful as this this means the new
@system-service system call filter group will get serious real-life
testing quickly.
This also switches from firing SIGSYS on unexpected syscalls to
returning EPERM. This would have probably been a better default anyway,
but it's hard to change that these days. When whitelisting system calls
SIGSYS is highly problematic as system calls that are newly introduced
to Linux become minefields for services otherwise.
Note that this enables a system call filter for udev for the first time,
and will block @clock, @mount and @swap from it. Some downstream
distributions might want to revert this locally if they want to permit
unsafe operations on udev rules, but in general this shiuld be mostly
safe, as we already set MountFlags=shared for udevd, hence at least
@mount won't change anything.
This adds a small service "systemd-portabled" and a matching client
"portablectl", which implement the "portable service" concept.
The daemon implements the actual operations, is PolicyKit-enabled and is
activated on demand with exit-on-idle.
Both the daemon and the client are an optional build artifact, enabled
by default rhough.
Unfortunately this needs a new binary to do the mount because there's just
too many special steps to outsource this to systemd-mount:
- EPERM needs to be treated specially
- UserRuntimeDir= setting must be obeyed
- SELinux label must be adjusted
This allows user@.service to be started independently of logind.
So 'systemctl start user@nnn' will start the user manager for user nnn.
Logind will start it too when the user logs in, and will stop it (unless
lingering is enabled) when the user logs out.
Fixes#7339.
This removes the UserTasksMax= setting in logind.conf. Instead, the generic
TasksMax= setting on the slice should be used. Instead of a transient unit we
use a drop-in to tweak the default definition of a .slice. It's better to use
the normal unit mechanisms instead of creating units on the fly. This will also
make it easier to start user@.service independently of logind, or set
additional settings like MemoryMax= for user slices.
The setting in logind is removed, because otherwise we would have two sources
of "truth": the slice on disk and the logind config. Instead of trying to
coordinate those two sources of configuration (and maintainer overrides to
both), let's just convert to the new one fully.
Right now now automatic transition mechanism is provided. logind will emit a
hint when it encounters the setting, but otherwise it will be ignored.
Fixes#2556.
Systems that have an accurate real-time clock may have an initial
unsynchronized time that is close enough to the synchronized time that
the final adjustment doesn't trigger a waking "clock set" event. Have
timesyncd touch a file in its runtime directory as a secondary signal
for synchronization. Continue to support the timerfd-based trigger as a
sufficient condition when the watchfile is not present.
Closes issue #8683
Files which are installed as-is (any .service and other unit files, .conf
files, .policy files, etc), are left as is. My assumption is that SPDX
identifiers are not yet that well known, so it's better to retain the
extended header to avoid any doubt.
I also kept any copyright lines. We can probably remove them, but it'd nice to
obtain explicit acks from all involved authors before doing that.
We still get the errors logged, but we don't fail the service. This
is better for users because rerunning tmpfiles-setup.service a second
time is dangerous (c.f. cd9f5b68ce).
Note that this only touches sd-tmpfiles-setup.service and
sd-tmpfiles-setup-dev.service. sd-tmpfiles-clean.service is as before.
https://bugzilla.redhat.com/show_bug.cgi?id=1539341
Use `systemctl --user --force exit` to implement the systemd-exit
user service.
This removes our dependence on an external `kill` binary and the
concerns about whether they recognize SIGRTMIN+n by name or what their
interpretation of SIGRTMIN is.
Tested: `systemctl --user start systemd-exit.service` kills the
`systemd --user` instance for my user.
Suspend to Hibernate is a new sleep method that invokes suspend
for a predefined period of time before automatically waking up
and hibernating the system.
It's similar to HybridSleep however there isn't a performance
impact on every suspend cycle.
It's intended to use with systems that may have a higher power
drain in their supported suspend states to prevent battery and
data loss over an extended suspend cycle.
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
CPU accounting has a too bad impact on performance to be enabled by
default. Therefore we should not delegate "cpu" for now.
OTOH since commit e0c46a7364, memory accounting
has been turned on for all units by default so it makes sense to delegate this
controller by default.
`-.mount` is placed in `system.slice`, and hence depends on it.
`-.mount` is always active and can never be stopped. Therefore the same
should be true of `system.slice`.
Synthesize it as perpetual (unless systemd is running as a user manager).
Notice we also drop `Before=slices.target` as unnecessary.
AFAICS the justification for `perpetual` is to provide extra protection
against unintentionally stopping every single service. So adding
system.slice to the perpetual units is perfectly consistent.
I don't expect this will (or can) fix any other problem. And the
`perpetual` protection probably isn't formal enough to spend much time
thinking about. I've just noticed this a couple of times, as something
that looks strange.
Might be a bit surprising that we have user.slice on-disk but not
system.slice, but I think it's ok. `systemctl status system.slice` will
still point you towards `man systemd.special`. The only detail is that the
system slice disables `DefaultDependencies`. If you're worrying about how
system shutdown works when you read `man systemd.slice`, I think it is not
too hard to guess that system.slice might do this:
> Only slice units involved with early boot
> or late system shutdown should disable this option
(Docs are great. I really appreciate the systemd ones).
On systems that only use resolved for name resolution, there are usecases that
require resolved to be started before sysinit target, such that network name
resolution is available before network-online/sysinit targets. For example,
cloud-init for some datasources hooks into the boot process ahead of sysinit
target and may need network name resolution at that point already.
systemd-resolved already starts pretty early in the process, thus starting it
slightly earlier should not have negative side effects.
However, this depends on resolved ability to connect to system DBus once that
is up.
Now that we can configure which controllers to delegate precisely, let's
limit wht we delegate to the user session: only "cpu" and "pids" as a
minimal baseline.
Fixes: #1715
An explicit --user switch is necessary because for the user@0.service instance
systemd-tmpfiles is running as root, and we need to distinguish that from
systemd-tmpfiles running in systemd-tmpfiles*.service.
Fixes#2208.
v2:
- restore "systemd-" prefix
- add systemd-tmpfiles-clean.{service,timer}, systemd-setup.service to
systemd-tmpfiles(8)
This makes sense from the point of view of the whole distribution:
if there are some specific files that have syntax problems, or unknown
users or groups, or use unsupported features, failing the whole service
is not useful.
In particular, services with tmpfiles --boot should not be started after boot.
The premise of --boot is that there are actions which are only safe to do once
during boot, because the state evolves later through other means and re-running
the boot-time setup would destroy it. If services with --boot fail in the
initial transaction, they would be re-run later on when a unit which
(indirectly) depends on them is started, causing problems.
Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1507501.
(If we had a mode where a service would at most run once, and would not be
started in subsequent transactions, that'd be a good additional safeguard.
Using ExecStart=-... is a bit like that, but it causes all failure to be
ignored, which is too big of a hammer.)
So far I avoided adding license headers to meson files, but they are pretty
big and important and should carry license headers like everything else.
I added my own copyright, even though other people modified those files too.
But this is mostly symbolic, so I hope that's OK.
There should be a way to turn this logic of, and DefaultDependencies=
appears to be the right option for that, hence let's downgrade this
dependency type from "implicit" to "default, and thus honour
DefaultDependencies=.
This also drops mount_get_fstype() as we only have a single user needing
this now.
A follow-up for #7076.
remote-cryptsetup-pre.target was designed as an active unit (that pulls in
network-online.target), the opposite of remote-fs-pre.target (a passive unit,
with individual provider services ordering itself before it and pulling it in,
for example iscsi.service and nfs-client.target).
To make remote-cryptsetup-pre.target really work, those services should be
ordered before it too. But this would require updates to all those services,
not just changes from systemd side.
But the requirements for remote-fs-pre.target and remote-cryptset-pre.target
are fairly similar (e.g. iscsi devices can certainly be used for both), so
let's reuse remote-fs-pre.target also for remote cryptsetup units. This loses
a bit of flexibility, but does away with the requirement for various provider
services to know about remote-cryptsetup-pre.target.
In the past we introduced this property just for tmp.mount. However on
todays systems usually there are many more tmpfs mounts. Most notably
mounts backing XDG_RUNTIME_DIR for each user.
Let's generalize what we already have for tmp.mount and implement the
ordering After=swap.target for all tmpfs based mounts.
This makes this target the same as remote-fs.target in this regard. In practice
it probably doesn't make that much difference, because all encrypted devices
that are part of remote-fs.target (marked with _netdev) will be used for mount
points, so they will be pulled in anyway individually, but with this change any
such device will be configured, even if it is not pulled by any other unit.
After the discussions around #7003 I think we should restore the
User=systemd-journal-gateway line for systemd-journal-gatewayd.service,
too, so that we continue to use the state user if it exists, and create
it as dynamic user only when it does not.
Note that undoes part of a change made after 234, i.e. a never released
change.
The configuration option was called -Dresolve, but the internal define
was …RESOLVED. This options governs more than just resolved itself, so
let's settle on the version without "d".
The advantage is that is the name is mispellt, cpp will warn us.
$ git grep -Ee "conf.set\('(HAVE|ENABLE)_" -l|xargs sed -r -i "s/conf.set\('(HAVE|ENABLE)_/conf.set10('\1_/"
$ git grep -Ee '#ifn?def (HAVE|ENABLE)' -l|xargs sed -r -i 's/#ifdef (HAVE|ENABLE)/#if \1/; s/#ifndef (HAVE|ENABLE)/#if ! \1/;'
$ git grep -Ee 'if.*defined\(HAVE' -l|xargs sed -i -r 's/defined\((HAVE_[A-Z0-9_]*)\)/\1/g'
$ git grep -Ee 'if.*defined\(ENABLE' -l|xargs sed -i -r 's/defined\((ENABLE_[A-Z0-9_]*)\)/\1/g'
+ manual changes to meson.build
squash! build-sys: use #if Y instead of #ifdef Y everywhere
v2:
- fix incorrect setting of HAVE_LIBIDN2
1. If we exited emergency mode immediately, we don't want to have an
irreversible stop job still running for syslog.socket. I _suspect_ that
can't happen, but let's not waste effort working out exactly why it's
impossible and not just very improbable.
2. Similarly, it seems undesirable to have rescue.service and
emergency.service both running with an open FD of /dev/console, for
however short a period.
Note this commit only changes how the code is expressed; it does not change
the existence of any dependency.
The `Conflicts=` was added in 3136ec90, "Stop syslog.socket when entering
emergency mode". The discussion in the issue #266 raised concerns that
this might be needed for other units, but failed to point out why
syslog.socket is special. The reason is that syslog.socket has
DefaultDepedencies=no, so it does not get Requires=sysinit.target like
other socket units do. But syslog.service does require sysinit.target,
among other things.
We don't have many socket, path, or timer units with
DefaultDependencies=no, and I don't think any of the triggered services
have such additional hard dependencies as syslog.service does.
It is much less confusing if we keep this `Conflicts=` in the same file as
the `DefaultDependencies=no` which made it necessary.
The original aim of this commit is that starting machines.target from the
rescue shell would not kill the rescue shell and lock you out of the
system.
This is similar to commit 6579a622, for the conflict between
sysinit.target and the _emergency_ shell. That particular commit
introduced an ordering cycle and will need to be reverted and/or
fixed. This one does not, because it does not need to introduce any new
dependencies.
The reason why this commit is allowable also has it's own merit:
machines.target was not marked as AllowIsolate. Also, the point of
containers is to not escape them... I don't think we want to promote
machines.target as a default target or similar; you would generally want
some system service to allow you to shut down the machine, for example. I
don't see this approach used in CoreOS, nor in Fedora Atomic Host; we are
missing any positive examples of its utility.
Requires=basic.target / After=basic.target can be removed for the same
reason.
This reverts commit f1e24a259c. Oops.
# systemctl emergency
Failed to start emergency.target: Transaction order is cyclic. See syste...
See system logs and 'systemctl status emergency.target' for details.
# systemctl status emergency.target
● emergency.target - Emergency Mode
Loaded: loaded (/usr/lib/systemd/system/emergency.target; static; vendor preset: disabled)
Active: inactive (dead) since Mon 2017-09-25 10:43:02 BST; 2h 42min ago
Docs: man:systemd.special(7)
systemd[1]: sysinit.target: Found dependency on sysinit.target/stop
sysinit.target: Unable to break cycle starting with sysinit.target/stop
network.target: Found ordering cycle on wpa_supplicant.service/stop
network.target: Found dependency on sysinit.target/stop
network.target: Found dependency on emergency.target/start
network.target: Found dependency on emergency.service/start
network.target: Found dependency on serial-getty@ttyS0.service/stop
network.target: Found dependency on systemd-user-sessions.service/stop
network.target: Found dependency on network.target/stop
network.target: Unable to break cycle starting with network.target/stop
IMO #6509 is ugly enough that we should aim to answer it. But it could
take some time to investigate, so let's re-open the issue as a first step.
Why
---
The advantage of this is that starting sysinit.target from the emergency
shell will no longer kill the emergency shell and lock you out of the
system. Our docs already claimed that emergency.target was useful for
"starting individual units in order to continue the boot process in steps".
This resolves#6509 for my purposes.
Remaining limitation
--------------------
Starting getty.target will still kill the shell, and if you don't have a
root password you will then be locked out at that point. This is relevant
to distributions which patch the sulogin system to permit logins when the
root password is locked. Both Debian and RedHat used to follow this
behaviour! Debian have been discussing what they could replace it with at
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=806852
So this doesn't quite achieve perfection, but I think it's a worthwhile
change. It should be easier to understand the logic now it doesn't have
such a big hole in it. Repairing the sysinit stage of the boot is the main
reason we have emergency.target. And as discussed in the issue,
sysinit.target gets pulled in implicitly as soon as any DefaultDependencies
service is activated.
How
---
sysinit.target only needs to conflict with emergency.target. It didn't
need to conflict with emergency.service as well. In theory the conflicts
are pointless, we could just change the dependency of sysinit.target on
local-fs.target from Wants to Requires. However, doing so would mean that
when local-fs fails, the screen is flooded with yellow [DEPEND] failures.
That would hinder the poor unfortunate admin, so let's not do that.
There is no additional ordering requirement against emergency. If the
failure happens, the job for sysinit will be cancelled instantly. We don't
need to worry about when sysinit.target and its dependents would be
stopped, because sysinit waits for local-fs before it starts.
emergency.target is still necessarily stopped once we reach sysinit
(you can't express a one-way conflict in pure unit directives).
This is largely cosmetic... though perhaps it symbolizes that you're no
longer in Emergency Mode if System Initialization is successful ;-).
As a secondary advantage, the getty's which conflict on rescue.service now
need to conflict on emergency.service as well. This makes the system more
uniform and simpler to understand.
The only other effect this should have is that
`systemctl start emergency.target` is now practically the same as
`systemctl start rescue.target`. The only units this command will stop are
the conflicting getty units. Neither of those commands should ever be
used. E.g. they will not stop the gdm.service unit on Fedora 26.
The pair is similar to remote-fs.target and remote-fs-pre.target. Any
cryptsetup devices which require network shall be ordered after
remote-cryptsetup-pre.target and before remote-cryptsetup.target.
Normally this happens automatically, but if it happened that both targets were
pulled in, even though there were no cryptsetup units, they could be started
in reverse order, which would be somewhat confusing. Add an explicit ordering
to avoid this potential issue.
This new target is a passive unit, hence it is supposed to be pulled in
to the transaction by the service that wants to block login on the
console (e.g. text version of initial-setup). Now both getty and
serial-getty are ordered after this target.
https://lists.freedesktop.org/archives/systemd-devel/2015-July/033754.html
and the same for hibernate.target and hybrid-sleep.target.
Tested with both sucessful and unsuccessful suspends. The result of the
start job was correct in both cases. Closes#6419 (a regression in v233
and v234).
> suspend is unsual for a target, because it has to stop itself once it's
> started. Otherwise you couldn't start it again, so you could only suspend
> once! Currently that's implemented using BindsTo=systemd-sleep.service.
> Meaning it pulls in systemd-sleep.service to do the actual suspend, and
> then de-activates afterwards. But the behaviour of BindsTo was changed
> recently (not without some issues during development) - maybe this bug
> is caused by poettering/systemd@631b676 which I think was added in
> release v233.
>
> sleep.target (see man systemd.special) has the same need, but it
> implements it differently. It simply has StopWhenUnneeded=yes.
This commit switches suspend.target etc. to the approach used by
sleep.target.
