SIGTERM should be considered a clean exit code for daemons (i.e. long-running
processes, as a daemon without SIGTERM handler may be shut down without issues
via SIGTERM still) while it should not be considered a clean exit code for
commands (i.e. short-running processes).
Let's add two different clean checking modes for this, and use the right one at
the appropriate places.
Fixes: #4275
Let's get rid of is_clean_exit_lsb(), let's move the logic for the special
handling of the two LSB exit codes into the sysv-generator by writing out
appropriate SuccessExitStatus= lines if the LSB header exists. This is not only
semantically more correct, bug also fixes a bug as the code in service.c that
chose between is_clean_exit_lsb() and is_clean_exit() based this check on
whether a native unit files was available for the unit. However, that check was
bogus since a long time, since the SysV generator was introduced and native
SysV script support was removed from PID 1, as in that case a unit file always
existed.
Show is documented to be program-parseable, and printing the warning about
about a non-existent unit, while useful for humans, broke a lot of scripts.
Restore previous behaviour of returning success and printing empty or useless
stuff for units which do not exist, and printing empty values for properties
which do not exists.
With SYSTEMD_LOG_LEVEL=debug, hints are printed, but the return value is
still 0.
This undoes parts of e33a06a and 3dced37b7 and fixes#3856.
We might consider adding an explicit switch to fail on missing units/properties
(e.g. --ensure-exists or similar), and make -P foobar equivalent to
--ensure-exists --property=foobar.
Prior to this commit, users could be given an unusable command to run if
they attempted to stop or start special services. For example:
$ systemctl stop -- -.mount
Failed to stop -.mount: Operation refused, unit -.mount may be \
requested by dependency only.
See system logs and 'systemctl status -.mount' for details.
$ systemctl status -.mount
systemctl: invalid option -- '.'
This adds a '--' to the example command in these situations.
Otherwise for example services that are failing on start and have Restart=on-failure
and bigger RestartSec systemctl status will return 0.
Fixes: #3864
systemd now returns an error when it is asked to perform disable on the
unit file path. In the past this was allowed, but systemd never really
considered an actual content of the [Install] section of the unit
file. Instead it performed disable on the unit name, i.e. purged all
symlinks pointing to the given unit file (undo of implicit link action
done by systemd when enable is called on the unit file path) and all
symlinks that have the same basename as the given unit file.
However, to notice that [Install] info of the file is not consulted one
must create additional symlinks manually. I argue that in most cases
users do not create such links. Let's be nice to our users and don't
break existing scripts that expect disable to work with the unit file
path.
Fixes#3706.
% valgrind --leak-check=full systemctl status multipathd.service --no-pager -n0
...
==431== 16 bytes in 2 blocks are definitely lost in loss record 1 of 2
==431== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299)
==431== by 0x534AF19: strdup (in /usr/lib64/libc-2.23.so)
==431== by 0x4E81AEE: free_and_strdup (string-util.c:794)
==431== by 0x4EF66C1: map_basic (bus-util.c:1030)
==431== by 0x4EF6A8E: bus_message_map_all_properties (bus-util.c:1153)
==431== by 0x120487: show_one (systemctl.c:4672)
==431== by 0x1218F3: show (systemctl.c:4990)
==431== by 0x4EC359E: dispatch_verb (verbs.c:92)
==431== by 0x12A3AE: systemctl_main (systemctl.c:7742)
==431== by 0x12B1A8: main (systemctl.c:8011)
==431==
==431== LEAK SUMMARY:
==431== definitely lost: 16 bytes in 2 blocks
This happens because map_basic() strdups the strings. Other code in systemctl
assigns strings to UnitStatusInfo without copying them, relying on the fact
that the message is longer lived than UnitStatusInfo. Add a helper function
that is similar to map_basic, but only accepts strings and does not copy them.
The alternative of continuing to use map_basic() but adding proper cleanup
to free fields in UnitStatusInfo seems less attractive because it'd require
changing a lot of code and doing a lot of more allocations for little gain.
(I put "leaking" in quotes, because systemctl is short lived anyway.)
Clarify that "systemctl enable" can operate either on unit names or on unit
file paths (also, adjust the --help text to clarify this). Say that "systemctl
enable" on unit file paths also links the unit into the search path.
Many other fixes.
This should improve the documentation to avoid further confusion around #3706.
Don't check inhibitors when operating remotely. The interactivity inhibitors
imply can#t be provided anyway, and the current code checks for local sessions
directly, via various sd_session_xyz() APIs, hence bypass it entirely if we
operate on remote systems.
Fixes: #3476
If we show both a control and a main PID for a service fix this line in the
output of "systemctl status":
Main PID: 19670 (sleep); : 19671 (sleep)
to become this:
Main PID: 19670 (sleep); Control PID: 19671 (sleep)
When unit has multiple condition list, systemctl is not showing which
conditions were failed. When user want to know which conditions were
failed, user has to check for each conditions.
So, show failed condition list also.
In https://github.com/systemd/systemd/issues/3543, we would open the pager
before starting ssh, and the pipe fd was "leaked" into the ssh child as the
stderr fd. Previous commit fixes bus-socket to nullify stderr before launching
the child, but it seems reasonable to also delay starting the pager.
If we are going to croak when trying to open the transport, it seems better
to do this before starting the pager.
This commit would also fix#3543 on its own.
If "systemctl -H" is used, let's make sure we first terminate the bus
connection, and only then close the pager. If done in this order ssh will get
an EOF on stdin (as we speak D-Bus through ssh's stdin/stdout), and then
terminate. This makes sure the standard error we were invoked on is released by
ssh, and only that makes sure we don't deadlock on the pager which waits for
all clients closing its input pipe.
(Similar fixes for the various other xyzctl tools that support both pagers and
-H)
Fixes: #3543
This reworks "systemctl status" and "systemctl show" a bit. It removes the
definition of the `property_info` structure, because we can simply reuse the
existing UnitStatusInfo type for that.
The "could not be found" message is now printed by show_one() itself (and not
its caller), so that it is shown regardless by who the function is called.
(This makes it necessary to pass the unit name to the function.)
This also adds all properties found to a set, and then checks if any of the
properties passed via "--property=" is mising in it, if so, a proper error is
generated.
Support for checking the PID file of a unit is removed, as this cannot be done
reasonably client side (since the systemd instance we are talking to might sit
on another host)
Replaces: #3411Fixes: #3425
Also see: #3504
... as well as halt/poweroff/kexec/suspend/hibernate/hybrid-sleep.
Running those commands will fail in user mode, but we try to set the wall
message first, which might even succeed for privileged users. Best to nip
the whole sequence in the bud.
https://github.com/systemd/systemd/pull/3453#issuecomment-225455156
Reloading or reexecuting PID 1 means the unit generators are rerun, which are
timed out at 90s. Make sure the method call asking for the reload is timed out
at twice that, so that the generators have 90s and the reload operation has 90s
too.
This reworks the daemon_reload() call in systemctl, and makes it exclusively
about reloading/reexecing. Previously it was used for other trivial method
calls too, which didn't really help readability. As the code paths are now
sufficiently different, split out the old code into a new function
trivial_method().
This call also does a similar change as
c8ad4efb277c3235d58789170af11bb3c847d655 but for the reload/reexec operation.
Fixes: #3353
For legacy commands such as /sbin/halt or /sbin/poweroff we support legacy
fallbacks that talk via traditional SysV way with PID 1 to issue the desired
operation. We do this on any kind of error if the primary method of operation
fails. When this is the case we suppress any error message that is normally
generated, in order to not confuse the user. When suppressing this log message,
don't suppress the original error code, because there's really no reason to.
systemctl --property doesn't validate if a requested property is valid or not,
and always returns with an exit code of 0, regardless of whether the requested
property exists or not.
How reproducible:
This works fine:
Id=multi-user.target
But put in a non-existing property:
Id=default.targets.service
Id=default.targets.service
0
[root@shou18lkvm8 ~]# systemctl show --property Id this.is.rubbish; echo $?
Id=this.is.rubbish.service
0
After:
sus@maximus bz-95593]$ ./systemctl show --property Id this.is.rubbish; echo $?
Can't display property this.is.rubbish. Unit this.is.rubbish.service does not
exist.
4
fixes#2295
