Previously we did operations like attach, trim or migrate only on the
controllers that were enabled for a specific unit. With this changes we
will now do them for all supproted controllers, and fall back to all
possible prefix paths if the specified paths do not exist.
This fixes issues if a controller is being disabled for a unit where it
was previously enabled, and makes sure that all processes stay as "far
down" the tree as groups exist.
Previously for an instantiated unit foo@bar.service we created a cgroup
foo@.service/foo@bar.service, in order to place all instances of the
same template inside the same subtree. As we now implicitly add all
instantiated units into one per-template slice we don't need this
complexity anymore, and instance units can map directly to the cgroups
of their full name.
During shutdown, when we try to clean up all remaining processes, the
kernel will fork new agents every time a cgroup runs empty. These
new processes cause delays in the final SIGTERM, SIGKILL logic.
Apart from that, this should also avoid that the kernel-forked binaries
cause unpredictably timed access to the filesystem which we might need to
unmount.
"Scope" units are very much like service units, however with the
difference that they are created from pre-existing processes, rather
than processes that systemd itself forks off. This means they are
generated programmatically via the bus API as transient units rather
than from static configuration read from disk. Also, they do not provide
execution-time parameters, as at the time systemd adds the processes to
the scope unit they already exist and the parameters cannot be applied
anymore.
The primary benefit of this new unit type is to create arbitrary cgroups
for worker-processes forked off an existing service.
This commit also adds a a new mode to "systemd-run" to run the specified
processes in a scope rather then a transient service.
Replace the very generic cgroup hookup with a much simpler one. With
this change only the high-level cgroup settings remain, the ability to
set arbitrary cgroup attributes is removed, so is support for adding
units to arbitrary cgroup controllers or setting arbitrary paths for
them (especially paths that are different for the various controllers).
This also introduces a new -.slice root slice, that is the parent of
system.slice and friends. This enables easy admin configuration of
root-level cgrouo properties.
This replaces DeviceDeny= by DevicePolicy=, and implicitly adds in
/dev/null, /dev/zero and friends if DeviceAllow= is used (unless this is
turned off by DevicePolicy=).
- This changes all logind cgroup objects to use slice objects rather
than fixed croup locations.
- logind can now collect minimal information about running
VMs/containers. As fixed cgroup locations can no longer be used we
need an entity that keeps track of machine cgroups in whatever slice
they might be located. Since logind already keeps track of users,
sessions and seats this is a trivial addition.
- nspawn will now register with logind and pass various bits of metadata
along. A new option "--slice=" has been added to place the container
in a specific slice.
- loginctl gained commands to list, introspect and terminate machines.
- user.slice and machine.slice will now be pulled in by logind.service,
since only logind.service requires this slice.
In order to prepare for the kernel cgroup rework, let's introduce a new
unit type to systemd, the "slice". Slices can be arranged in a tree and
are useful to partition resources freely and hierarchally by the user.
Each service unit can now be assigned to one of these slices, and later
on login users and machines may too.
Slices translate pretty directly to the cgroup hierarchy, and the
various objects can be assigned to any of the slices in the tree.
I'm assuming that it's fine if a _const_ or _pure_ function
calls assert. It is assumed that the assert won't trigger,
and even if it does, it can only trigger on the first call
with a given set of parameters, and we don't care if the
compiler moves the order of calls.
clang emits warnings about unused attribute _saved_errno_, which drown
out other—potentially useful—warnings. gcc documentation is not exactly
verbose about the effects of __attribute__((unused)) on variables, but
let's assume that it works if the unit test passes.
Session objects will now get the .session suffix, user objects the .user
suffix, nspawn containers the .nspawn suffix.
This also changes the user cgroups to be named after the numeric UID
rather than the username, since this allows us the parse these paths
standalone without requiring access to the cgroup file system.
This also changes the mapping of instanced units to cgroups. Instead of
mapping foo@bar.service to the cgroup path /user/foo@.service/bar we
will now map it to /user/foo@.service/foo@bar.service, in order to
ensure that all our objects are properly suffixed in the tree.
As discussed with Dan Berrange it's a good idea to suffix all objects in
the cgroup tree with ".something", so that when the system is
partitioned using a resource management tool we can drop objects of
different types into the same partition directory without generate
namespace conflicts.
We'l add this to the Pax Control Group document as soon as write access
to the fdo wiki is restored.
Containers will now carry a label (normally derived from the root
directory name, but configurable by the user), and the container's root
cgroup is /machine/<label>. This label is called "machine name", and can
cover both containers and VMs (as soon as libvirt also makes use of
/machine/).
libsystemd-login can be used to query the machine name from a process.
This patch also includes numerous clean-ups for the cgroup code.