docs: Add libvirt misc doc

A new dumping ground for libvirt'y things that can't easily be
categorized or placed elsewhere.

Change-Id: I6999b9d66e12e1df7970aff6ce63e5323de6be45
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Related-Bug: #1843542
This commit is contained in:
Stephen Finucane 2021-02-22 18:58:41 +00:00
parent 76549775fe
commit 9dfac32959
4 changed files with 141 additions and 109 deletions

View File

@ -384,18 +384,6 @@ See `the KVM documentation
information on these limitations.
Guest agent support
-------------------
Use guest agents to enable optional access between compute nodes and guests
through a socket, using the QMP protocol.
To enable this feature, you must set ``hw_qemu_guest_agent=yes`` as a metadata
parameter on the image you wish to use to create the guest-agent-capable
instances from. You can explicitly disable the feature by setting
``hw_qemu_guest_agent=no`` in the image metadata.
KVM performance tweaks
----------------------

View File

@ -120,6 +120,7 @@ instance for these kind of workloads.
managing-resource-providers
resource-limits
cpu-models
libvirt-misc
Additional guides

View File

@ -0,0 +1,140 @@
======================
Other libvirt features
======================
The libvirt driver supports a large number of additional features that don't
warrant their own section. These are gathered here.
Guest agent support
-------------------
Guest agents enable optional access between compute nodes and guests through a
socket, using the QMP protocol.
To enable this feature, you must set ``hw_qemu_guest_agent=yes`` as a metadata
parameter on the image you wish to use to create the guest-agent-capable
instances from. You can explicitly disable the feature by setting
``hw_qemu_guest_agent=no`` in the image metadata.
.. _extra-specs-watchdog-behavior:
Watchdog behavior
-----------------
.. versionchanged:: 15.0.0 (Ocata)
Add support for the ``disabled`` option.
A virtual watchdog device can be used to keep an eye on the guest server and
carry out a configured action if the server hangs. The watchdog uses the
i6300esb device (emulating a PCI Intel 6300ESB). Watchdog behavior can be
configured using the :nova:extra-spec:`hw:watchdog_action` flavor extra spec or
equivalent image metadata property. If neither the extra spec not the image
metadata property are specified, the watchdog is disabled.
For example, to enable the watchdog and configure it to forcefully reset the
guest in the event of a hang, run:
.. code-block:: console
$ openstack flavor set $FLAVOR --property hw:watchdog_action=reset
.. note::
Watchdog behavior set using the image metadata property will override
behavior set using the flavor extra spec.
.. _extra-specs-random-number-generator:
Random number generator
-----------------------
.. versionchanged:: 21.0.0 (Ussuri)
Random number generators are now enabled by default for instances.
Operating systems require good sources of entropy for things like cryptographic
software. If a random-number generator device has been added to the instance
through its image properties, the device can be enabled and configured using
the :nova:extra-spec:`hw_rng:allowed`, :nova:extra-spec:`hw_rng:rate_bytes` and
:nova:extra-spec:`hw_rng:rate_period` flavor extra specs.
To configure for example a byte rate of 5 bytes per period and a period of 1000
mSec (1 second), run:
.. code-block:: console
$ openstack flavor set $FLAVOR \
--property hw_rng:rate_bytes=5 \
--property hw_rng:rate_period=1000
Alternatively, to disable the random number generator, run:
.. code-block:: console
$ openstack flavor set $FLAVOR --property hw_rng:allowed=false
The presence of separate byte rate and rate period configurables is
intentional. As noted in the `QEMU docs`__, a smaller rate and larger period
minimizes the opportunity for malicious guests to starve other guests of
entropy but at the cost of responsiveness. Conversely, larger rates and smaller
periods will increase the burst rate but at the potential cost of warping
resource consumption in favour of a greedy guest.
.. __: https://wiki.qemu.org/Features/VirtIORNG#Effect_of_the_period_parameter
.. _extra-specs-performance-monitoring-unit:
Performance Monitoring Unit (vPMU)
----------------------------------
.. versionadded:: 20.0.0 (Train)
If nova is deployed with the libvirt virt driver and
:oslo.config:option:`libvirt.virt_type` is set to ``qemu`` or ``kvm``, a
virtual performance monitoring unit (vPMU) can be enabled or disabled for an
instance using the :nova:extra-spec:`hw:pmu` flavor extra spec or ``hw_pmu``
image metadata property.
If the vPMU is not explicitly enabled or disabled via
the flavor or image, its presence is left to QEMU to decide.
For example, to explicitly disable the vPMU, run:
.. code-block:: console
$ openstack flavor set FLAVOR-NAME --property hw:pmu=false
The vPMU is used by tools like ``perf`` in the guest to provide more accurate
information for profiling application and monitoring guest performance.
For :doc:`real time </admin/real-time>` workloads, the emulation of a vPMU can
introduce additional latency which would be undesirable. If the telemetry it
provides is not required, the vPMU can be disabled. For most workloads the
default of unset (enabled) will be correct.
.. _extra-specs-hiding-hypervisor-signature:
Hiding hypervisor signature
---------------------------
.. versionadded:: 18.0.0 (Rocky)
.. versionchanged:: 21.0.0 (Ussuri)
Prior to the Ussuri release, this was called ``hide_hypervisor_id``. An
alias is provided to provide backwards compatibility.
Some hypervisors add a signature to their guests. While the presence of the
signature can enable some paravirtualization features on the guest, it can also
have the effect of preventing some drivers from loading. You can hide this
signature by setting the :nova:extra-spec:`hw:hide_hypervisor_id` to true.
For example, to hide your signature from the guest OS, run:
.. code:: console
$ openstack flavor set $FLAVOR --property hw:hide_hypervisor_id=true

View File

@ -115,103 +115,6 @@ Hardware video RAM
for more information on how this is used to set the ``videoRamSizeInKB`` attribute with
the vmware driver.
.. _extra-specs-watchdog-behavior:
Watchdog behavior
For the libvirt driver, you can enable and set the behavior of a virtual
hardware watchdog device for each flavor. Watchdog devices keep an eye on the
guest server, and carry out the configured action, if the server hangs. The
watchdog uses the i6300esb device (emulating a PCI Intel 6300ESB). If
``hw:watchdog_action`` is not specified, the watchdog is disabled.
To set the behavior, use:
.. code-block:: console
$ openstack flavor set FLAVOR-NAME --property hw:watchdog_action=ACTION
Valid ACTION values are:
- ``disabled``: (default) The device is not attached.
- ``reset``: Forcefully reset the guest.
- ``poweroff``: Forcefully power off the guest.
- ``pause``: Pause the guest.
- ``none``: Only enable the watchdog; do nothing if the server hangs.
.. note::
Watchdog behavior set using a specific image's properties will override
behavior set using flavors.
.. _extra-specs-random-number-generator:
Random-number generator
If a random-number generator device has been added to the instance through
its image properties, the device can be enabled and configured using:
.. code-block:: console
$ openstack flavor set FLAVOR-NAME \
--property hw_rng:allowed=True \
--property hw_rng:rate_bytes=RATE-BYTES \
--property hw_rng:rate_period=RATE-PERIOD
Where:
- RATE-BYTES: (integer) Allowed amount of bytes that the guest can read from
the host's entropy per period.
- RATE-PERIOD: (integer) Duration of the read period in milliseconds.
.. _extra-specs-performance-monitoring-unit:
Performance Monitoring Unit (vPMU)
If nova is deployed with the libvirt virt driver and
:oslo.config:option:`libvirt.virt_type` is set to ``qemu`` or ``kvm``, a
vPMU can be enabled or disabled for an instance using the ``hw:pmu``
extra_spec or the ``hw_pmu`` image property.
The supported values are ``True`` or ``False``. If the vPMU is not
explicitly enabled or disabled via the flavor or image, its presence is left
to QEMU to decide.
.. code-block:: console
$ openstack flavor set FLAVOR-NAME --property hw:pmu=True|False
The vPMU is used by tools like ``perf`` in the guest to provide more accurate
information for profiling application and monitoring guest performance.
For realtime workloads, the emulation of a vPMU can introduce additional
latency which may be undesirable. If the telemetry it provides is not
required, such workloads should set ``hw:pmu=False``. For most workloads
the default of unset or enabling the vPMU ``hw:pmu=True`` will be correct.
.. _extra-specs-hiding-hypervisor-signature:
Hiding hypervisor signature
Some hypervisors add a signature to their guests. While the presence
of the signature can enable some paravirtualization features on the
guest, it can also have the effect of preventing some drivers from
loading. Hiding the signature by setting this property to true may
allow such drivers to load and work.
.. note::
As of the 18.0.0 Rocky release, this is only supported by the libvirt
driver.
Prior to the 21.0.0 Ussuri release, this was called
``hide_hypervisor_id``. An alias is provided to provide backwards
compatibility.
.. code:: console
$ openstack flavor set FLAVOR-NAME \
--property hw:hide_hypervisor_id=VALUE
Where:
- VALUE: (string) 'true' or 'false'. 'false' is equivalent to the
property not existing.
.. _extra-specs-secure-boot:
Secure Boot