Two things here:
1. The API guide was missing the hyper-v driver which supports
the suspend operation. Rather than hard-code a list of supported
drivers in the doc, this change just links to the entry in the
feature support matrix.
2. The supported hypervisors mention in the API reference is removed
because the end user using the API should not need to know or care
what backend hypervisor on which their server is running. They can
either suspend or not, but we don't need to mention the supporting
drivers for that in the API reference.
Change-Id: Ib76779a8e34b2c68b0f4af190f71576180360d0f
Related-Bug: #1815403
While we do not have an automated fix for bug 1849479 this provides
a troubleshooting document for working around that issue where
allocations from a server that was evacuated from a down host need
to be cleaned up manually in order to delete the resource provider
and associated compute node/service.
In general this is also a useful guide for linking up the various
resources and terms in nova and how they are reflected in placement
with the relevant commands which is probably something we should
do more of in our docs.
Change-Id: I120e1ddd7946a371888bfc890b5979f2e19288cd
Related-Bug: #1829479
The description of "SUSPENDED" server status was misguiding. Rewording
it to make it more accurate.
Change-Id: Ie93b3b38c2000f7e9caa3ca89dea4ec04ed15067
Closes-Bug: #1815403
Add a section to the support matrix for image caching
(``has_imagecache`` virt driver capability).
Change-Id: I9147c5ea6b276b4fe18a981f4360844009bd3d95
Partial-Bug: #1847302
Blueprint image-precache-support added a conf section called
[image_cache], so it makes sense to move all the existing image
cache-related conf options into it.
Old:
[DEFAULT]image_cache_manager_interval
[DEFAULT]image_cache_subdirectory_name
[DEFAULT]remove_unused_base_images
[DEFAULT]remove_unused_original_minimum_age_seconds
[libvirt]remove_unused_resized_minimum_age_seconds
New:
[image_cache]manager_interval
[image_cache]subdirectory_name
[image_cache]remove_unused_base_images
[image_cache]remove_unused_original_minimum_age_seconds
[image_cache]remove_unused_resized_minimum_age_seconds
Change-Id: I3c49825ac0d70152b6c8ee4c8ca01546265f4b80
Partial-Bug: #1847302
The only kwarg passed to resize() is the auto_disk_config
kwarg during resize (not cold migrate). This expands it out
to a named kwarg to resolve the TODO. The same is done with
_check_auto_disk_config and as a result there is a hit to
rebuild as well, but the same TODO exists on rebuild() to use
named kwargs but that can be dealt with separately.
Change-Id: Ide8eb9e09d22f20165474d499ef0524aefc67854
This is a follow up to [1] to make the API behave consistently
by always asynchronously casting to conductor during resize
and cold migration regardless of same-cell or cross-cell
migration.
From the end user point of view, not much changes besides
the possibility of some exceptions occurring during scheduling
which would have resulted in a 400 BadRequest error.
The user still gets a 202 response, must poll the server status
until the server goes to VERIFY_RESIZE status or times out, and
can check the instance actions if the resize/migrate fails.
The specific errors that can occur are not really an API contract
and as such end user applications should not be building logic
around, for example, getting a NoValidHost error. It should be
noted, however, that by default non-admin users cannot see
the instance action event traceback that would contain the
error, e.g. NoValidHost.
The only exception types removed from handling in the API are
(1) AllocationMoveFailed which can be raised when the conductor
MigrationTask runs replace_allocation_with_migration and
(2) NoValidHost when the scheduler is called to select destinations.
Because of this, quite a few functional negative tests have to be
adjusted since the API no longer returns a 400 for NoValidHost and
other errors that can happen during scheduling.
Finally, the do_cast kwarg is left on the conductor API method since
the compute service calls it during same-cell reschedule as a
synchronous RPC call and has error handling if rescheduling in
conductor fails.
[1] I098f91d8c498e5a85266e193ad37c08aca4792b2
Change-Id: I711e56bcb4b72605253fa63be230a68e03e45b84
Before [1] this could be raised from the API resize()
method if getting a RequestSpec failed and a target host
was specified for cold migration. Since that change the
usage of the exception was removed so we can remove it
altogether since only unit test code is using it.
[1] I34ffaf285718059b55f90e812b57f1e11d566c6f
Change-Id: I19db48bd03855d1a1edbeff5adf15a28abcb5d92
The TODOs were added back in the Queens/Pike timeframe [1][2]
but at this point there probably isn't much value in resolving
those TODOs by adding a skip_filters kwarg to the scheduler
especially since [3] changed the method to not support nested
resource provider allocations so the minimal duplication with
what we do in the scheduler in the non-force evacuate/live migrate
cases is sufficient.
[1] Ie63a4798d420c39815e294843e02ab6473cfded2
[2] I6590f0eda4ec4996543ad40d8c2640b83fc3dd9d
[3] I7cbd5d9fb875ebf72995362e0b6693492ce32051
Change-Id: I3e599147f95337477c9573b517feee67e0ae37e4
The Selection object returned by select_destinations has a
compute node UUID in it so we don't have to look up the
compute node object by host/nodename when reverting allocations
during live migration.
Change-Id: I0156cda8543f847d50e16683e1eb29fbdd556d27
The server group functional tests are doing time.sleep(10) in order
to make sure a stopped compute service is considered "down" by the nova
compute API.
Instead of sleeping, we can set the service as "forced_down" to get the
desired "down" compute service status and avoid unnecessary delays in
these tests.
Unnecessary service start() calls are also removed in this change. They
appear at the end of tests and services are started during each test
setUp() and killed during each test tearDown() via the ServiceFixture.
Closes-Bug: #1783565
Change-Id: I74f64b68e4b33ee0f8c45fdc5f570c7e12e05d3b
With Ib62ac0b692eb92a2ed364ec9f486ded05def39ad and the
get_inventory method gone nothing uses this so we can
remove it now.
Change-Id: I3f55e09641465279b8b92551a2302219fe6fc5ca
In Train [1] we deprecated support for compute drivers
that did not implement the update_provider_tree method.
That compat code is now removed along with the get_inventory
method definition and (most) references to it.
As a result there are more things we can remove but those
will come in separate changes.
[1] I1eae47bce08f6292d38e893a2122289bcd6f4b58
Change-Id: Ib62ac0b692eb92a2ed364ec9f486ded05def39ad
Matt Riedemann