systemd-soft-reboot.service — Userspace reboot operation
systemd-soft-reboot.service
systemd-soft-reboot.service
is a system service that is pulled in by
soft-reboot.target
and is responsible for performing a userspace-only reboot
operation. When invoked, it will send the SIGTERM
signal to any processes left
running (but does not wait for the processes to exit), and follow up with SIGKILL
.
If the /run/nextroot/
directory exists (which may be a regular directory, a
directory mount point or a symlink to either) then it will switch the file system root to it. It then
reexecutes the service manager off the (possibly now new) root file system, which will enqueue a new boot
transaction as in a normal reboot.
Such a userspace-only reboot operation permits updating or resetting the entirety of userspace with minimal downtime, as the reboot operation does not transition through:
The second phase of regular shutdown, as implemented by systemd-shutdown(8).
The third phase of regular shutdown, i.e. the return to the initrd context.
The hardware reboot operation.
The firmware initialization.
The boot loader initialization.
The kernel initialization.
The initrd initialization.
However this form of reboot comes with drawbacks as well:
The OS update remains incomplete, as the kernel is not reset and continues running.
Kernel settings (such as /proc/sys/
settings, a.k.a. "sysctl", or
/sys/
settings) are not reset.
These limitations may be addressed by various means, which are outside of the scope of this
documentation, such as kernel live-patching and sufficiently comprehensive
/etc/sysctl.d/
files.
Various runtime OS resources can passed from a system runtime to the next, through the userspace reboot operation. Specifically:
File descriptors placed in the file descriptor store of services that remain active
until the very end are passed to the next boot, where they are placed in the file descriptor store of
the same unit. For this to work, units must declare DefaultDependencies=no
(and
avoid a manual Conflicts=shutdown.target
or similar) to ensure they are not
terminated as usual during the system shutdown operation. Alternatively, use
FileDescriptorStorePreserve=
to allow the file descriptor store to remain pinned
even when the unit is down. See
systemd.service(5) for
details about the file descriptor store.
Similar to this, file descriptors associated with .socket
units
remain open (and connectible) if the units are not stopped during the transition. (Achieved by
DefaultDependencies=no
.)
The /run/
file system remains mounted and populated and may be
used to pass state information between such userspace reboot cycles.
Service processes may continue to run over the transition, past soft-reboot and into
the next session, if they are placed in services that remain active until the very end of shutdown
(which again is achieved via DefaultDependencies=no
). They must also be set up to
avoid being killed by the aforementioned SIGTERM
and SIGKILL
via SurviveFinalKillSignal=yes
, and also be configured to avoid being stopped on
isolate via IgnoreOnIsolate=yes
. They also have to be configured to be stopped on
normal shutdown, reboot and maintenance mode. Finally, they have to be ordered after
basic.target
to ensure correct ordering on boot. Note that in case any new or
custom units are used to isolate to, or that implement an equivalent shutdown functionality, they will
also have to be configured manually for correct ordering and conflicting. For example:
[Unit] Description=My Surviving Service SurviveFinalKillSignal=yes IgnoreOnIsolate=yes DefaultDependencies=no After=basic.target Conflicts=reboot.target kexec.target poweroff.target halt.target rescue.target emergency.target Before=shutdown.target rescue.target emergency.target [Service] Type=oneshot ExecStart=sleep infinity
On top of the above, templated units also need a configuration file for their slice, as
they by default use a slice named after the non-templated part of the unit. For example, for a
foo@test.service
instance, a system-foo.slice
unit may be
added with the following content:
[Unit] SurviveFinalKillSignal=yes IgnoreOnIsolate=yes DefaultDependencies=no
File system mounts may remain mounted during the transition, and complex storage
attached, if configured to remain until the very end of the shutdown process. (Also achieved via
DefaultDependencies=no
, and by avoiding
Conflicts=umount.target
)
If the unit publishes a service over D-Bus, the connection needs to be re-established after soft-reboot as the D-Bus broker will be stopped and then started again. When using the sd-bus library this can be achieved by adapting the following example.
/* SPDX-License-Identifier: MIT-0 */ /* A D-Bus service that automatically reconnects when the system bus is * restarted. * * Compile with 'cc sd_bus_service_reconnect.c $(pkg-config --libs --cflags libsystemd)' * * To allow the program to take ownership of the name 'org.freedesktop.ReconnectExample', * add the following as /etc/dbus-1/system.d/org.freedesktop.ReconnectExample.conf * and then reload the broker with 'systemctl reload dbus': <?xml version="1.0"?> <!--*-nxml-*--> <!DOCTYPE busconfig PUBLIC "-//freedesktop//DTD D-BUS Bus Configuration 1.0//EN" "http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd"> <busconfig> <policy user="root"> <allow own="org.freedesktop.ReconnectExample"/> <allow send_destination="org.freedesktop.ReconnectExample"/> <allow receive_sender="org.freedesktop.ReconnectExample"/> </policy> <policy context="default"> <allow send_destination="org.freedesktop.ReconnectExample"/> <allow receive_sender="org.freedesktop.ReconnectExample"/> </policy> </busconfig> * * To get the property via busctl: * * $ busctl --user get-property org.freedesktop.ReconnectExample \ * /org/freedesktop/ReconnectExample \ * org.freedesktop.ReconnectExample \ * Example * s "example" */ #include <errno.h> #include <stdio.h> #include <stdlib.h> #include <systemd/sd-bus.h> #define _cleanup_(f) __attribute__((cleanup(f))) static int log_error(int r, const char *str) { fprintf(stderr, "%s failed: %s\n", str, strerror(-r)); return r; } typedef struct object { const char *example; sd_bus **bus; sd_event **event; } object; static int property_get( sd_bus *bus, const char *path, const char *interface, const char *property, sd_bus_message *reply, void *userdata, sd_bus_error *error) { object *o = userdata; if (strcmp(property, "Example") == 0) return sd_bus_message_append(reply, "s", o->example); return sd_bus_error_setf(error, SD_BUS_ERROR_UNKNOWN_PROPERTY, "Unknown property '%s'", property); } /* https://www.freedesktop.org/software/systemd/man/sd_bus_add_object.html */ static const sd_bus_vtable vtable[] = { SD_BUS_VTABLE_START(0), SD_BUS_PROPERTY( "Example", "s", property_get, 0, SD_BUS_VTABLE_PROPERTY_CONST), SD_BUS_VTABLE_END }; static int setup(object *o); static int on_disconnect(sd_bus_message *message, void *userdata, sd_bus_error *ret_error) { int r; r = setup((object *)userdata); if (r < 0) { object *o = userdata; r = sd_event_exit(*o->event, r); if (r < 0) return log_error(r, "sd_event_exit()"); } return 1; } /* Ensure the event loop exits with a clear error if acquiring the well-known * service name fails */ static int request_name_callback(sd_bus_message *m, void *userdata, sd_bus_error *ret_error) { int r; if (!sd_bus_message_is_method_error(m, NULL)) return 1; const sd_bus_error *error = sd_bus_message_get_error(m); if (sd_bus_error_has_names(error, SD_BUS_ERROR_TIMEOUT, SD_BUS_ERROR_NO_REPLY)) return 1; /* The bus is not available, try again later */ fprintf(stderr, "Failed to request name: %s\n", error->message); object *o = userdata; r = sd_event_exit(*o->event, -sd_bus_error_get_errno(error)); if (r < 0) return log_error(r, "sd_event_exit()"); return 1; } static int setup(object *o) { int r; /* If we are reconnecting, then the bus object needs to be closed, detached * from the event loop and recreated. * https://www.freedesktop.org/software/systemd/man/sd_bus_detach_event.html * https://www.freedesktop.org/software/systemd/man/sd_bus_close_unref.html */ if (*o->bus) { r = sd_bus_detach_event(*o->bus); if (r < 0) return log_error(r, "sd_bus_detach_event()"); *o->bus = sd_bus_close_unref(*o->bus); } /* Set up a new bus object for the system bus, configure it to wait for D-Bus * to be available instead of failing if it is not, and start it. All the * following operations are asynchronous and will not block waiting for D-Bus * to be available. * https://www.freedesktop.org/software/systemd/man/sd_bus_new.html * https://www.freedesktop.org/software/systemd/man/sd_bus_set_address.html * https://www.freedesktop.org/software/systemd/man/sd_bus_set_bus_client.html * https://www.freedesktop.org/software/systemd/man/sd_bus_negotiate_creds.html * https://www.freedesktop.org/software/systemd/man/sd_bus_set_watch_bind.html * https://www.freedesktop.org/software/systemd/man/sd_bus_set_connected_signal.html * https://www.freedesktop.org/software/systemd/man/sd_bus_start.html */ r = sd_bus_new(o->bus); if (r < 0) return log_error(r, "sd_bus_new()"); r = sd_bus_set_address(*o->bus, "unix:path=/run/dbus/system_bus_socket"); if (r < 0) return log_error(r, "sd_bus_set_address()"); r = sd_bus_set_bus_client(*o->bus, 1); if (r < 0) return log_error(r, "sd_bus_set_bus_client()"); r = sd_bus_negotiate_creds(*o->bus, 1, SD_BUS_CREDS_UID|SD_BUS_CREDS_EUID|SD_BUS_CREDS_EFFECTIVE_CAPS); if (r < 0) return log_error(r, "sd_bus_negotiate_creds()"); r = sd_bus_set_watch_bind(*o->bus, 1); if (r < 0) return log_error(r, "sd_bus_set_watch_bind()"); r = sd_bus_start(*o->bus); if (r < 0) return log_error(r, "sd_bus_start()"); /* Publish an interface on the bus, specifying our well-known object access * path and public interface name. * https://www.freedesktop.org/software/systemd/man/sd_bus_add_object.html * https://dbus.freedesktop.org/doc/dbus-tutorial.html */ r = sd_bus_add_object_vtable(*o->bus, NULL, "/org/freedesktop/ReconnectExample", "org.freedesktop.ReconnectExample", vtable, o); if (r < 0) return log_error(r, "sd_bus_add_object_vtable()"); /* By default the service is only assigned an ephemeral name. Also add a * well-known one, so that clients know whom to call. This needs to be * asynchronous, as D-Bus might not be yet available. The callback will check * whether the error is expected or not, in case it fails. * https://www.freedesktop.org/software/systemd/man/sd_bus_request_name.html */ r = sd_bus_request_name_async(*o->bus, NULL, "org.freedesktop.ReconnectExample", 0, request_name_callback, o); if (r < 0) return log_error(r, "sd_bus_request_name_async()"); /* When D-Bus is disconnected this callback will be invoked, which will set up * the connection again. This needs to be asynchronous, as D-Bus might not yet * be available. * https://www.freedesktop.org/software/systemd/man/sd_bus_match_signal_async.html */ r = sd_bus_match_signal_async(*o->bus, NULL, "org.freedesktop.DBus.Local", NULL, "org.freedesktop.DBus.Local", "Disconnected", on_disconnect, NULL, o); if (r < 0) return log_error(r, "sd_bus_match_signal_async()"); /* Attach the bus object to the event loop so that calls and signals are * processed. * https://www.freedesktop.org/software/systemd/man/sd_bus_attach_event.html */ r = sd_bus_attach_event(*o->bus, *o->event, 0); if (r < 0) return log_error(r, "sd_bus_attach_event()"); return 0; } int main(int argc, char **argv) { /* The bus should be relinquished before the program terminates. The cleanup * attribute allows us to do it nicely and cleanly whenever we exit the block. */ _cleanup_(sd_bus_flush_close_unrefp) sd_bus *bus = NULL; _cleanup_(sd_event_unrefp) sd_event *event = NULL; object o = { .example = "example", .bus = &bus, .event = &event, }; int r; /* Create an event loop data structure, with default parameters. * https://www.freedesktop.org/software/systemd/man/sd_event_default.html */ r = sd_event_default(&event); if (r < 0) return log_error(r, "sd_event_default()"); /* By default the event loop will terminate when all sources have disappeared, * so we have to keep it 'occupied'. Register signal handling to do so. * https://www.freedesktop.org/software/systemd/man/sd_event_add_signal.html */ r = sd_event_add_signal(event, NULL, SIGINT|SD_EVENT_SIGNAL_PROCMASK, NULL, NULL); if (r < 0) return log_error(r, "sd_event_add_signal(SIGINT)"); r = sd_event_add_signal(event, NULL, SIGTERM|SD_EVENT_SIGNAL_PROCMASK, NULL, NULL); if (r < 0) return log_error(r, "sd_event_add_signal(SIGTERM)"); r = setup(&o); if (r < 0) return EXIT_FAILURE; /* Enter the main loop, it will exit only on sigint/sigterm. * https://www.freedesktop.org/software/systemd/man/sd_event_loop.html */ r = sd_event_loop(event); if (r < 0) return log_error(r, "sd_event_loop()"); /* https://www.freedesktop.org/software/systemd/man/sd_bus_release_name.html */ r = sd_bus_release_name(bus, "org.freedesktop.ReconnectExample"); if (r < 0) return log_error(r, "sd_bus_release_name()"); return 0; }
Even though passing resources from one soft reboot cycle to the next is possible this way, we
strongly suggest to use this functionality sparingly only, as it creates a more fragile system as
resources from different versions of the OS and applications might be mixed with unforeseen
consequences. In particular it's recommended to avoid allowing processes to survive
the soft reboot operation, as this means code updates will necessarily be incomplete, and processes
typically pin various other resources (such as the file system they are backed by), thus increasing
memory usage (as two versions of the OS/application/file system might be kept in memory). Leaving
processes running during a soft-reboot operation requires disconnecting the service comprehensively from
the rest of the OS, i.e. minimizing IPC and reducing sharing of resources with the rest of the OS. A
possible mechanism to achieve this is the concept of Portable Services, but make sure no resource from the
host's OS filesystems is pinned via BindPaths=
or similar unit settings, otherwise the
old, originating filesystem will remain mounted as long as the unit is running.
Note that because
systemd-shutdown(8) is
not executed, the executables in /usr/lib/systemd/system-shutdown/
are not executed
either.
Note that systemd-soft-reboot.service
(and related units) should never be
executed directly. Instead, trigger system shutdown with a command such as "systemctl
soft-reboot
".
Note that if a new root file system has been set up on "/run/nextroot/
", a
soft-reboot will be performed when the reboot command is
invoked.