This changes does two things to the admin scheduler configuration
docs:
1. Notes the limitation from bug 1802111 for the older
AggregateMultiTenancyIsolation filter and mentions that
starting in Rocky, using tenant isolation with placement
is better.
2. Notes that when isolating tenants via placement, the metadata
key "filter_tenant_id" can be suffixed to overcome the limitation
in bug 1802111.
Change-Id: I792c5df01b7cbc46c8363e261bc7422b09180e56
Closes-Bug: #1802111
Recently, the _ThreadingEvent class in oslo.service was removed [1] and
our unit test patching is preventing us from moving to a newer version
of oslo.service [2].
We have patching of the _ThreadingEvent.wait method to bypass the sleep
time in the looping call of RetryDecorator, which adds several seconds
to the run time of unit tests.
This changes things to use the new SleepFixture from oslo.service
instead.
Depends-On: https://review.openstack.org/616371
[1] I62e9f1a7cde8846be368fbec58b8e0825ce02079
[2] https://review.openstack.org/615676
Change-Id: I45dd7602068eb0ce1331cfefd5a0cf6418bc8e88
We need to completely remove the mocks for oslo.service private members
in order to decouple nova from the requirements updates and be able to
update the constraint for oslo.service to a new version that has a
fixture for managing the sleep call being mocked (applied in the
following patch).
Change-Id: I0bbd2d7f9d6eb13d97587d867ef4d651809a7dd4
Signed-off-by: Doug Hellmann <doug@doughellmann.com>
- This change tries to address an edge case discovered
when running placement under mod_wsgi where
if the placement wsgi application is re-initialize the db_api
configure method attempts to reconfigure a started transaction
factory.
- This since oslo.db transaction factories do not support
reconfiguration at runtime this result in an exception being
raised preventing reloading of the Placement API without
restarting apache to force mod_wsgi to recreate the
python interpreter.
- This change introduces a run once decorator to allow annotating
functions that should only be executed once for the lifetime fo
an interpreter.
- This change applies the run_once decorator to the db_api configure
method, to suppress the attempt to reconfigure the current
TransactionFactory on application reload.
Co-Authored-By: Balazs Gibizer <balazs.gibizer@ericsson.com>
Closes-Bug: #1799246
Related-Bug: #1784155
Change-Id: I704196711d30c1124e713ac31111a8ea6fa2f1ba
Replace the 'NOTE' and 'TODO' words with
the note and todo directives in the Compute API guide.
Improve some format in the guide.
e.g. Fix unnecessary bold characters.
Change-Id: I7e9d6b8a5228886fd5de03d12482481d2b652666
Closes-Bug: #1558459
Things have changed here on Walton's Mountain since LazyLoad was
introduced [1]. It seems to have been created to avoid a circular
import, but as this patch should attest, that's no longer an issue.
Why change this now, besides removing weird and complicated code?
Because a subsequent patch needs to access a *property* of the report
client from the compute manager. As written, LazyLoad only lets you get
at *methods* (as functools.partial). There are other ways to solve that
while preserving the deferred import, but if this works, it's the better
solution.
[1] Ie5732baf9709cd0cb951eae4638910372c79e5f1
Change-Id: I1f97d00fb7633f173370ed6787c9a71ecd8106d5
With this change, [compute]resource_provider_association_refresh is
allowed to be zero, which will disable refreshing resource provider
traits and aggregates.
Inventories are still refreshed in a different code path.
A subsequent patch will be submitted to allow manual refresh by sending
SIGHUP to the compute process.
Change-Id: Iec33e656491848b26686fbf6fb5db4a4c94b9ea8
When creating a snapshot of a server using the nova API, failure occurs
if the image contains the metadata property "version". This was due to
the way that the GlanceClientWrapper.call() function signature was
structured.
This patch forces all client positional args to be passed as a named
"args" argument to the call() function and all client named args to be
pass as a named "kwargs" argument to the call() function. This
eliminates any argument name-shadowing that previously caused issues.
Closes-bug: #1717547
Change-Id: I3ed3303309fe2a25c0043fd206f36bada4b3b8f9
UEFI support in the VMware driver has been added with commit fc0c6d2.
This patch fixes the support matrix to reflect this.
Change-Id: I8b08e11ae4dd7f1101758b29ae3424d790b26ed1
The new RequestSpec.resource_requests stores a list of RequestGroup
objects each representing a granular resource request group requesting
resource for the given instance.
This will be used in a later patch to communicate the Neutron port's
resource request to the scheduler. Also this can be used later to
refactor flavor-based resource request handling.
This new field is intentionally not persisted to the DB for two
reasons:
* the port's resource request is already persisted in neutron
so this would be a data duplication that could lead to data
inconsistencies
* this field is only used to carry information from the the
nova-conductor to the nova-scheduler for each select_destinations()
request
blueprint: bandwidth-resource-provider
Change-Id: I53e5debcffd6de2b3a2ff838e7f5da33fa1a82b8
Later patches will introduce a field in RequestSpec using this type as
the field type to store the resource requests coming from outside of
Nova like the bandwidth request coming from the Neutron ports.
This patch refactors the usage of placement.lib.RequestGroup. Until now
this class was used both by placement and nova services and they used
it only as a util class. However after this series the nova services
would like to use such a class via RPC which requires an OVO. This
patch makes sure that the new OVO is used by nova and the old plain
object is used by placement. This way placement is not forced to use
an OVO where no OVO functionality is required.
The minimum required version of oslo.versionedobjects is updated to
1.33.3 to include the fix for bug 1771804.
Change-Id: I46c97d2641d9685ef59771314665a17a5236097d
blueprint: bandwidth-resource-provider
With [1] now a thing, we should strive to keep Tempest's default for
scheduler_enabled_filters the same as Nova's enabled_filters. This
patch adds a comment to that effect.
[1] https://review.openstack.org/#/c/570207/
Related-bug: 1628443
Change-Id: I0a02acb831d00b472b9d80b4f4205a7f1476feb5
after the changed[1], the nubmers of service components on
controller node should be two rather than three.
[1]: https://review.openstack.org/#/c/604277/
Change-Id: Iada43eb7f36f946d1713b20a50ebdeb8c69d0545
Closes-Bug: #1801444
We currently pass all members of the disk_info dict to libvirt_info
when generating libvirt config. A subsequent change adds another key
to this dict. Rather than add a fourth argument to libvirt_info, this
pre-emptive change just passes the dict.
Part of bp local-disk-serial-numbers
Change-Id: I28c5bc23c0ea60d64153472d8937965f60f907c4
The tests for some libvirt volume were following the same pattern of
passing invalid disk_info data to get_config. They seemed to expect
connection_info['data'] to contain the same data as disk_info, which
has never been true.
Part of bp local-disk-serial-numbers
Change-Id: I0fc5ebc9eeae8feb644adaa96537e69521acec78
Small modification to correct spelling mistake.
Change-Id: I4bf378e5316ecc48f66eefae4f45d5a505adc305
Signed-off-by: Nguyen Hai Truong <truongnh@vn.fujitsu.com>
The PowerVM patch to implement the update_provider_tree method (in a
fashion backward compatible with the existing get_available_resource
method) [1] was in flight at the same time as a
bp reshape-provider-tree patch introducing the 'allocations' kwarg [2]
and another patch which provided a ComputeDriver helper method for
calculating the reserved value for the DISK_GB resource from config [3].
This brings PowerVM's update_provider_tree method up to scratch with
those changes.
[1] I96da858e6a63e3339a009fd17ee0d8939b364510
[2] Ic062446e5c620c89aec3065b34bcdc6bf5966275
[3] I72c83a95dabd581998470edb9543079acb6536a5
Change-Id: I89ad36f19672368a1f795e1f29c5af6368ccfeec
This patch modifies the existing behavior of nova service-list which
ignores a down cell to returning a partial construct consisting of
the binary and host info obtained from the API host_mappings table
for the compute services in the down cells. This is enabled by passing
a new enough microversion, which is added in a later patch in this
series.
Related to blueprint handling-down-cell
Change-Id: I9cf5e4eb7f70c495001fd064c060d7d5b3dd5bff
This patch sets the stage for modifying the behavior of nova show
which currently gives a 500 when the cell in which the instance
lives is down. The new behavior will return a partial construct
consisting of uuid, project_id, created_at from instance_mappings
table and user_id, flavor, image_ref and availability_zone info
from request_specs table. Note that the rest of the keys will be
missing. This behavior will be enabled by passing a new enough
microversion, handling for which is introduced later in this series.
Related to blueprint handling-down-cell
Change-Id: Iaea1cb4ed93bb98f451de4f993106d7891ca3682
Scatter-gather utility returns a raised_exception_sentinel for
all kinds of exceptions that are caught and often times there
maybe situations where we may have to handle the different types
of exceptions differently. To facilitate that, it might be more
useful to return the Exception object itself instead of the dummy
raised_exception_sentinel so that based on the result's exception
type we can handle them differently.
Related to blueprint handling-down-cell
Change-Id: I861b223ee46b0f0a31f646a4b45f8a02410253cf
This patch augments the existing behavior of nova list which
ignores a down cell to returning a partial construct consisting of
uuid, project_id and created_at fields obtained from the API
instance_mappings table for the instances in the down cell by
introducing a new API microversion. This way when the user
gets partial results for certain records, he/she will know
that that particular cell is unreachable. Note that the other
keys will be missing from the API response. Also note that this
is not externally poke-able until the microversion and API change
later in this series.
Related to blueprint handling-down-cell
Change-Id: Ie9a9aed0676d469e5180d4c2d2631cf339335b2d
InstanceMappingList.get_by_cell_and_project() is being changed to
get_not_deleted_by_cell_and_project() by which we change the method to
filter out not queued_for_delete instance_mappings and also for
the method to accept cell_uuid as the input instead of the cell_id.
This also adds an optional limit parameter. This is more useful than
the ealier approach since it would increase the performance slightly
by not having to query for the cell_mappings explicitly to get the
cell ids from the cell uuids obtained from scatter gather utility.
Note that this method was added relatively recently and hence is not
used anywhere which is why it should be fine to change this, keeping
in mind the intention of having added it in the first place.
Change-Id: I5df9851cce0b6e635c74fd34f5be9c4906dd7406
Related to blueprint handling-down-cell
Zuul