nova/api-guide/source/port_with_resource_request.rst
Balazs Gibizer fee98b7147 Allow migrating server with port resource request
Now that the nova code supports such a migration this patch removes the
check from the API that rejected the operation.

Note that in the spec [1] we agreed not to introduce new microversion
for this change but treat it as a bugfix. The current change also makes
it possible to accept the migration of these servers with _any_
microversion.

[1] https://specs.openstack.org/openstack/nova-specs/specs/train/approved/support-move-ops-with-qos-ports.html#rest-api-impact

Change-Id: I4bda569cc7c247e83219276282724c77e760ddcd
blueprint: support-move-ops-with-qos-ports
2019-09-12 10:32:51 -04:00

32 lines
1.4 KiB
ReStructuredText

=================================
Using ports with resource request
=================================
Starting from microversion 2.72 nova supports creating servers with neutron
ports having resource request visible as a admin-only port attribute
``resource_request``. For example a neutron port has resource request if it has
a QoS minimum bandwidth rule attached. Deleting such servers or detaching such
ports works since Stein version of nova without requiring any specific
microversion.
However the following API operations are still not supported in nova:
* Creating servers with neutron networks having QoS minimum bandwidth rule is
not supported. The user needs to pre-create the port in that neutron network
and create the server with the pre-created port.
* Attaching Neutron ports and networks having QoS minimum bandwidth rule is not
supported.
Also the following API operations are not supported in the 19.0.0 (Stein)
version of nova:
* Moving (resizing, migrating, live-migrating, evacuating, unshelving after
shelve offload) servers with ports having resource request is not yet
supported.
As of 20.0.0 (Train), nova supports cold migrating servers with neutron ports
having resource requests if both the source and destination compute services
are upgraded to 20.0.0 (Train) and the ``[upgrade_levels]/compute``
configuration does not prevent the computes from using the latest RPC version.