The QEMU and LXC drivers make use of the Linux "Control Groups" facility for applying resource management to their virtual machines and containers.
The control groups filesystem supports multiple "controllers". By default
the init system (such as systemd) should mount all controllers compiled
into the kernel at /sys/fs/cgroup/$CONTROLLER-NAME. Libvirt
will never attempt to mount any controllers itself, merely detect where
they are mounted.
The QEMU driver is capable of using the cpuset,
cpu, memory, blkio and
devices controllers. None of them are compulsory.
If any controller is not mounted, the resource management APIs
which use it will cease to operate. It is possible to explicitly
turn off use of a controller, even when mounted, via the
/etc/libvirt/qemu.conf configuration file.
The LXC driver is capable of using the cpuset,
cpu, cpuset, freezer,
memory, blkio and devices
controllers. The cpuset, devices
and memory controllers are compulsory. Without
them mounted, no containers can be started. If any of the
other controllers are not mounted, the resource management APIs
which use them will cease to operate.
As of libvirt 1.0.5 or later, the cgroups layout created by libvirt has been
simplified, in order to facilitate the setup of resource control policies by
administrators / management applications. The layout is based on the concepts of
"partitions" and "consumers". Each virtual machine or container is a consumer,
and has a corresponding cgroup named $VMNAME.libvirt-{qemu,lxc}.
Each consumer is associated with exactly one partition, which also have a
corresponding cgroup usually named $PARTNAME.partition. The
exceptions to this naming rule are the three top level default partitions,
named /system (for system services), /user (for
user login sessions) and /machine (for virtual machines and
containers). By default every consumer will of course be associated with
the /machine partition. This leads to a hierarchy that looks
like
$ROOT
|
+- system
| |
| +- libvirtd.service
|
+- machine
|
+- vm1.libvirt-qemu
| |
| +- emulator
| +- vcpu0
| +- vcpu1
|
+- vm2.libvirt-qemu
| |
| +- emulator
| +- vcpu0
| +- vcpu1
|
+- vm3.libvirt-qemu
| |
| +- emulator
| +- vcpu0
| +- vcpu1
|
+- container1.libvirt-lxc
|
+- container2.libvirt-lxc
|
+- container3.libvirt-lxc
The default cgroups layout ensures that, when there is contention for CPU time, it is shared equally between system services, user sessions and virtual machines / containers. This prevents virtual machines from locking the administrator out of the host, or impacting execution of system services. Conversely, when there is no contention from system services / user sessions, it is possible for virtual machines to fully utilize the host CPUs.
If there is a need to apply resource constraints to groups of
virtual machines or containers, then the single default
partition /machine may not be sufficiently
flexible. The administrator may wish to sub-divide the
default partition, for example into "testing" and "production"
partitions, and then assign each guest to a specific
sub-partition. This is achieved via a small element addition
to the guest domain XML config, just below the main domain
element
...
<resource>
<partition>/machine/production</partition>
</resource>
...
Libvirt will not auto-create the cgroups directory to back this partition. In the future, libvirt / virsh will provide APIs / commands to create custom partitions, but currently this is left as an exercise for the administrator. For example, given the XML config above, the admin would need to create a cgroup named '/machine/production.partition'
# cd /sys/fs/cgroup
# for i in blkio cpu,cpuacct cpuset devices freezer memory net_cls perf_event
do
mkdir $i/machine/production.partition
done
# for i in cpuset.cpus cpuset.mems
do
cat cpuset/machine/$i > cpuset/machine/production.partition/$i
done
Note: the cgroups directory created as a ".partition" suffix, but the XML config does not require this suffix.
Note: the ability to place guests in custom partitions is only available with libvirt >= 1.0.5, using the new cgroup layout. The legacy cgroups layout described later did not support customization per guest.
Since libvirt aims to provide an API which is portable across hypervisors, the concept of cgroups is not exposed directly in the API or XML configuration. It is considered to be an internal implementation detail. Instead libvirt provides a set of APIs for applying resource controls, which are then mapped to corresponding cgroup tunables
Parameters from the "cpu" controller are exposed via the
schedinfo command in virsh.
# virsh schedinfo demo Scheduler : posix cpu_shares : 1024 vcpu_period : 100000 vcpu_quota : -1 emulator_period: 100000 emulator_quota : -1
Parameters from the "blkio" controller are exposed via the
bkliotune command in virsh.
# virsh blkiotune demo weight : 500 device_weight :
Parameters from the "memory" controller are exposed via the
memtune command in virsh.
# virsh memtune demo
hard_limit : 580192
soft_limit : unlimited
swap_hard_limit: unlimited
The net_cls is not currently used. Instead traffic
filter policies are set directly against individual virtual
network interfaces.
Prior to libvirt 1.0.5, the cgroups layout created by libvirt was different
from that described above, and did not allow for administrator customization.
Libvirt used a fixed, 3-level hierarchy libvirt/{qemu,lxc}/$VMNAME
which was rooted at the point in the hierarchy where libvirtd itself was
located. So if libvirtd was placed at /system/libvirtd.service
by systemd, the groups for each virtual machine / container would be located
at /system/libvirtd.service/libvirt/{qemu,lxc}/$VMNAME. In addition
to this, the QEMU drivers further child groups for each vCPU thread and the
emulator thread(s). This leads to a hierarchy that looked like
$ROOT
|
+- system
|
+- libvirtd.service
|
+- libvirt
|
+- qemu
| |
| +- vm1
| | |
| | +- emulator
| | +- vcpu0
| | +- vcpu1
| |
| +- vm2
| | |
| | +- emulator
| | +- vcpu0
| | +- vcpu1
| |
| +- vm3
| |
| +- emulator
| +- vcpu0
| +- vcpu1
|
+- lxc
|
+- container1
|
+- container2
|
+- container3
Although current releases are much improved, historically the use of deep hierarchies has had a significant negative impact on the kernel scalability. The legacy libvirt cgroups layout highlighted these problems, to the detriment of the performance of virtual machines and containers.