When a service deactivates and is then automatically restarted via
Restart= we currently quickly transition through
SERVICE_DEAD/SERVICE_FAILED. Which is weird given it's not the
normal ("permanent") dead/failed state, but a transitory one we
immediately leave from again. We do this so that software that looks for
failures/successes can take notice, even if we restart as a consequence
of the deactivation.
Let's clean this up a bit: let's introduce two new states:
SERVICE_DEAD_BEFORE_AUTO_RESTART and SERVICE_FAILED_BEFORE_AUTO_RESTART
that are used for the transitory states. Both the SERVICE_DEAD and
SERVICE_DEAD_BEFORE_AUTO_RESTART will map to the high-level
UNIT_INACTIVE state though. (and similar for the respective failed
states). This means the high-level state machine won't change by this,
only the low-level one.
This clearly seperates the substates, which makes the state engine
cleaner, and allows clients to follow precisely whether we are in a
transitory dead/failed state, or a permanent one, by looking at the
service substate. Moreover it allows us to remove the 'n_keep_fd_store'
which so far we used to ensure the fdstore was not released during this
transitory dead/failed state but only during the permanent one. Since we
can now distinguish these states properly we can just use that.
This has been bugging me for a while. Let's clean this up.
Note that the unit restart logic is already nicely covered in the
testsiute, hence this adds no new tests for that.
And yes, this could be considered a compat break, but sofar we took the
liberty to make changes to the low-level state machine (i.e. SERVICE_xyz
states, sometimes called "substates") without considering this a bad
breakage – the high-level state machine (i.e. UNIT_xyz states) should
be considered API that cannot be changed.
- Drop Netdev= as it was removed in mkosi
- Always install python-psutil in the final image (required for networkd tests)
- Always Install python-pytest in the final image (required for ukify tests)
- Use the narrow glob for all centos python packages
- Drop the networkd mkosi config files (the default image can be used instead)
- Use ".conf" as the mkosi config file suffix everywhere
- Copy src/ to /root/src in the final image and set gdb substitute path in
.gdbinit to make gdb work properly
- ACLs are not set on generated directories anymore by default, so
we enable them explictly now so that when running unprivileged mkosi,
the user running mkosi can remove all generated files and directories.
- We don't explicitly set QemuHeadless= anymore as the option was removed
and made the default.
- We set the loglevel= kernel cmdline argument explicitly now as mkosi
doesn't set it by default anymore.
Let's open a file descriptor to the root directory and perform all
path operations using that file descriptor. On top of that, let's
make sure we pin the directory containing the file we want to work
on and use atomic copies and writes everywhere we can..
LOG_SET_PREFIX() sets a logging prefix for the current block. The
prefix is prepended to every logging message in the block, followed
by ": ". If a logging prefix is already configured, it is overridden
for the duration of the block, after which it is restored.
A use case for this macro is when we're operating on an image or
directory (using --root or --image). We can use LOG_SET_PREFIX() to
prefix all logging messages with the directory or image that we're
operating on.
This removes "sleep" invocations, and makes the notify access testcase a
lot more robust to runtime jitter. We use a pair of fifos in the fs to
sync instead.
Also various other improvoements, including comments.
(Also removes the unnecessary "no-qemu" restriction)
Unit names can be 255 characters long, not 256.
We first say "name prefix" and then continue with "unit prefix".
Confusing. Couldn't figure out which term is better hence settled on
"unit name prefix".
Let's skip going through all locally available time zones if we're built
with slow tests disabled, as that's quite slow, but do at least one test
with the UTC zone.
We shouldn't report that the file is empty if the stating fails. Let's do the
same as in other places, and just ignore the error and let the subsequent
operation fail.
This reverts commit 42a467b552.
We need to skip -ENOENT when loading udev rules because new files with rules
may be added or removed at any time, and the loading of rules is triggered
asynchronously. Even though the window is fairly narrow, udev shouldn't throw
an error if a rules file is removed.
Older stubs are either not stripped, causing their total size to be
unaligned because of an unaligned symbol table at the end, or stripped,
causing the raw data pointers and sizes to be unaligned because strip
does not follow the PE spec correctly when stripping. Let's add
workarounds for both issues, so that we can use ukify with older stubs
as well.
The reason why get_process_cmdline() is so complicated is that we
need to escape and quote arguments for building a single result
string.
That's necessary when we want to log or print the command line.
However, when we want to parse the command line, it is not necessary
that the result is a single string, but can be strv.
This will be used when we parse the command line.
sysext DDI cannot carry an os-release file, but have to carry
an extension-release file. But so far, this was only used to
match the sysext DDI with the base DDI/rootdir. It is also
useful to describe the sysext DDI itself, just like we do in
os-release.
So document that the same fields used in os-release can also
be added to an extension-release, with the 'SYSEXT_' prefix,
and in that case they are understood to define the sysext DDI
itself, rather than for matching purposes.
This is useful to identify log messages with metadata from the images
they run on. Look for ID/VERSION_ID/IMAGE_ID/IMAGE_VERSION/BUILD_ID,
with a SYSEXT_ prefix if we are looking at an extension, and append via
LogExtraFields= as respectively PORTABLE_NAME_AND_VERSION= in case of a
single image. In case of extensions, append as PORTABLE_ROOT_NAME_AND_VERSION=
for the base and one PORTABLE_EXTENSION_AND_VERSION= for each extension.
Example with a base and two extensions, with the unit coming from the
first extension:
[Service]
RootImage=/home/bluca/git/systemd/base.raw
Environment=PORTABLE=app0.raw
BindReadOnlyPaths=/etc/os-release:/run/host/os-release
LogExtraFields=PORTABLE=app0.raw
Environment=PORTABLE_ROOT=base.raw
LogExtraFields=PORTABLE_ROOT=base.raw
LogExtraFields=PORTABLE_ROOT_NAME_AND_VERSION=debian_10
ExtensionImages=/home/bluca/git/systemd/app0.raw
LogExtraFields=PORTABLE_EXTENSION=app0.raw
LogExtraFields=PORTABLE_EXTENSION_NAME_AND_VERSION=app_0
ExtensionImages=/home/bluca/git/systemd/app1.raw
LogExtraFields=PORTABLE_EXTENSION=app1.raw
LogExtraFields=PORTABLE_EXTENSION_NAME_AND_VERSION=app_1
When a portable service uses extensions, we use the 'main' image name
(the one where the unit was found in) as PORTABLE=. It is useful to
also list all the images actually used at runtime, as they might
contain libraries and so on.
Use PORTABLE_ROOT= for the image/directory that is used as RootImage=
or RootDirectory=, and PORTABLE_EXTENSION= for the image/directory that
is used as ExtensionImages= or ExtensionDirectories=.
Note that these new fields are only added if extensions are used,
there's no change for single-DDI portables.
Example with a base and two extensions, with the unit coming from the
first extension:
[Service]
RootImage=/home/bluca/git/systemd/base.raw
Environment=PORTABLE=app0.raw
BindReadOnlyPaths=/etc/os-release:/run/host/os-release
LogExtraFields=PORTABLE=app0.raw
LogExtraFields=PORTABLE_ROOT=base.raw
ExtensionImages=/home/bluca/git/systemd/app0.raw
LogExtraFields=PORTABLE_EXTENSION=app0.raw
ExtensionImages=/home/bluca/git/systemd/app1.raw
LogExtraFields=PORTABLE_EXTENSION=app1.raw
take_fdopen_unlocked invalidates the FD in the PortableMetadata object,
so it cannot be used later. Use parse_env_file_fd instead which is non
destructive.
Let's add a dash of colour to separate our own status info from the the
status info supplied by the service.
(I wanted to make this italics, but apparently popular terminal
emulators don't support that, such as xterm)
The fdstore might pin a non-trivial amount of resources. Let's hence
display for services that enable it how many entries there are and what
the size limit is.
When using an old systemctl on a new PID 1 the "startup" memory limit
props are not initialized and currently be shown in status output as
zero, even though there's code to suppress the output in that case. Alas
it doesn't work, because the relevant fields are not marked as
"unset"... Fix that.
Follow-up for: 53fda560dc
Lennart Poettering
Zbigniew Jędrzejewski-Szmek
Daan De Meyer
Luca Boccassi
Dmitry V. Levin
Yu Watanabe
OMOJOLA JOSHUA DAMILOLA
Franck Bui
Frantisek Sumsal