If there is no api reference for some placement API method
tox -e placement-api-ref will fail now.
Change-Id: Ifefedcd51f1f8d5e741ef6d02b7a0879a185e05f
Due to the scheduler report client on the compute service
not being aware of shared storage providers in Placement
for a given compute node, we can't claim support for
shared storage pool modeling in Pike. This adds a release
note to be clear about the lack of that functionality.
Change-Id: I0692a4fdf9a19889c5d708fb9a6a562c914963f9
Related-Bug: #1707256
Our release notes are read by a lot of people and we somehow need
to summarize the top main changes that are in that release for Nova.
Audience is targeted to be general and not really dev-oriented.
Co-Authored-By: Matt Riedemann <mriedem.os@gmail.com>
Change-Id: I65cfb6e22597edfeda687fc2f0c5ae5e098aa950
This provides a more detailed, but still high level, explanation
of how the FilterScheduler is using allocation candidates to pick
hosts during scheduling, and how the allocations are handled during
a move operation, including a resize to the same host.
Ideally this content will live somewhere else in the devref, or
would be updated in the spec for blueprint placement-claims, but
for now this should suffice for Pike release notes that can point
at this upgrade doc.
Change-Id: I274c684f829bc310ebb17faabf498d36f4ceea0c
This adds a test that shows we clean up allocations for an instance
that was local deleted while the compute host was down, when the
compute host comes back up.
There might still be a problem if the compute host is never brought
back up, as allocations will still exist for the instance and show
up as usage during usage queries to placement.
Related-Bug: #1679750
Change-Id: Ia68a5a69783963b063264edde84006973bb77ceb
This adds information about where the consoleauth service and console
proxies should run with multiple cells in Cells v2.
Change-Id: Ib817581dfe0c32d3888c242166e3daa7b954320a
This reorganizes the pieces that are for end users into a convenient
chunk up front in the main index page.
Part of bp: doc-migration
Change-Id: Ic2fcabbf3faa531fccb19cffa2564f005142e463
A large chunk of the release note for this is duplicated with
the same reno from e5080c7330
but the formatting and wording is slightly different.
This fixes them to look similar for the duplicate text.
Change-Id: Idd6149ae85ac2724633b80e938c4c1bf981b772b
Avoid instance info cache update if the instance is deleted,
as the deleted instance's info cache is already deleted in
db layer and it will report InstanceInfoCacheNotFound
exception and lead to a few exception logs in compute.
Please note this is different to regular info cache and
lead to InstanceInfoCacheNotFound exception, it's in the
case when you first create the info cache after instance
delete and InstanceInfo is also deleted.
Change-Id: I860e9e7c7ef458722135a21c6c5745f5519c56c4
Closes-Bug: 1618822
Import the following documents from the admin guide [1]:
- cli-manage-flavors.rst
- cli-nova-evacuate.rst
- cli-nova-manage-projects-security.rst
- cli-nova-manage-services.rst
- cli-nova-numa-libvirt.rst
- cli-nova-specify-host.rst
- cli-set-quotas.rst
- cli-set-compute-quotas.rst
- cli-os-migrate.rst
- cli-os-migrate-cfg-ssh.rst
- ts-no-emulator-x86-64.rst
- ts-multipath-warn.rst
- ts-failed-connect-vol-FC-SAN.rst
- ts-failed-attach-vol-no-sysfsutils.rst
- ts-failed-attach-vol-after-detach.rst
- ts-vol-attach-miss-sg-scan.rst
These were missed in the original import [2]. All 'ts-' files are merged
into a single 'support-compute' document, which is already in-tree.
[1] https://github.com/openstack/openstack-manuals/tree/stable/ocata/doc/admin-guide/source
[2] https://review.openstack.org/#/c/477497/
Change-Id: Ifa0039e270e54ea2fb58ab18ce6724e5e8e061a1
Closes-Bug: #1708666
In an attempt to make the main navigation sidebar not be visual mud
(and really confusing) create a reference sub page that explains all
the references.
Part of bp: doc-migration
Change-Id: I005fc24e49487631d7fe73cb498c61619bcaac9d
Placement took over the role of the CoreFilter, RamFilter and DiskFilter
from the FilterScheduler. Therefore if placement returns no allocation
candidates for a request then scheduling should be stopped as this means
there is not enough VCPU, MEMORY_MB, or DISK_GB available in any compute
node for the request.
Change-Id: If20a20e5cce7ab490998643e32556a1016646b07
Closes-Bug: #1708637
This fixes all the >= 2 hit 404s from 2017-08-07 with
redirects. nova-cert is being referenced somewhere, in that case we
provide a 410 because that's never coming back.
Part of bp: doc-migration
Change-Id: I313e4a22dc2ad3faa890de864d00943b802e8b30
The ironic driver uses ironicclient to get the list of nodes from the
Ironic service. If that service is not running, or something otherwise
prevents the ironicclient's call from succeeding, and unhandled
exception will occur. This patch adds a generic exception handler for
this call, logging an error and continuing.
Co-Authored-By: Lucas Alvares Gomes <lucasagomes@gmail.com>
Closes-Bug: #1706772
Change-Id: I4ebebcf221dfe29e2aa125f5956aec10108f8fbe
When a multiple-instance creation is requested by the user, we create a
single RequestSpec per instance (each of them having the corresponding
instance UUID) but we keep track of how many concurrent instances were
created at once by updating a field named 'num_instances' and we persist
it.
Unfortunately, due to Ifc5cf482209e4f6f4e3e39b24389bd3563d86444 we now
lookup that field in the scheduler for knowing how many allocations to
do. Since a move operation will only pass a single instance UUID to
migrate to the scheduler, there is a discrepancy that can lead to an
ugly IndexError. Defaulting that loop to be what is passed over RPC
unless we don't have it and then we default back to what we know from
the RequestSpec record.
Change-Id: If7da79356174be57481ef246618221e3b2ff8200
Closes-Bug: #1708961
XenServer 6.0 has long since been replaced and the bug was fixed in 6.1
[1]. No need to keep this legacy script around.
[1] https://bugs.launchpad.net/nova/+bug/1103158/comments/1
Change-Id: I8c41e88281aaf4250859622142a2f03f4452e1be
This test tries to resize to a flavor that requests more VCPU than
what the compute hosts has available and expect the resize to fail.
However placement returns no allocation candidate and the scheduler
falls back to the legacy filtering. As CoreFilter is not enabled the
filtering result in hosts selected without enough VCPU resource.
This is a test for an open bug therefore it asserts the wrong behavior
but it has the proper assert commented out.
Related-bug: #1708637
Change-Id: Ibd7f695d51b1f8497ca492968c7e5c69ada0a85c
Andrey Volkov