Libvirt has "domainSetUserPassword" callback and virtuozzo driver
has an implementation for it. So in this patch we allow to use this
functionality for virtuozzo hypervisor.
Change-Id: Ia398afadfd9fd9544c5d843338ab25c0930d9f74
Implements: blueprint virtuozzo-instance-admin-password
The description of "Inject guest networking config" in
support-matrix is misleading. Correct it.
In fact, If we config static ip,
guest os can not access metadata service to get ip config.
Closes-Bug: #1612913
Change-Id: I1892e5273a77516dd8e1525e6064fb95fa65e4d3
Adapt "nova rescue" code to support Virtuozzo ploop disks.
In rescue mode we use static mount point /mnt/rescue for original filesystem.
As far as ploop disks are in fact directories
we use rmtree function to delete rescue disks in unrescue procedure.
Implements: blueprint virtuozzo-rescue-support
Change-Id: I07b9e3a8b726635294395b0292fc3592a1b0f918
Adapt "nova resize" code to support Virtuozzo ploop disks.
As far as ploop disks are in fact directories we add '-r' argument
to all utilities that deal with instance' disks such as cp, rsync and scp.
Thus we copy disks universally whether they are folders or files.
Also using "prl_disk_tool" instead of "qemu-img" is better for ploop images
because it resizes guest filesystem as well.
We can't resize disks from guest OS in containers,
because they are not allowed to write directly to block device.
ploop tool can resize partition table and internal filesystem,
but only for container's disks. Such disks must have only one partition
with ext filesystem.
prl_disk_tool can resize disks with internal filesystems
and doesn't require any special layout so it can resize disks
for virtual machines. So it's better to use this tool instead of ploop.
Also we make compute.filters more strict
We call "ploop" only with "restore-descriptor" argument
And we set disk size in megabytes for prl_disk_tool
Co-Authored-By: Dmitry Guryanov <dguryanov@parallels.com>
Depends-On: I04c4379459c2fc1fd4801ec2aad53d0f6053b6d6
Change-Id: I38dbf73beb01fe1939ddca63fbfedbec1dc3c826
Implements: blueprint virtuozzo-instance-resize-support
HyperV: Nova serial console access support [1] has been merged
successfully. Update support matrix accordingly.
[1] https://review.openstack.org/145004
Change-Id: Ie6792e91c5c6c24d4af448605e4bb7d245bf41a8
Closes-Bug: #1584199
The root disk image is moved to a separate disk slot while the
rescue image will take it's place. If the instance requires it,
a temporary config drive is created as well.
Unrescuing the instance will move the root disk image back in
place, removing temporary images.
DocImpact
Implements: blueprint hyper-v-rescue
Change-Id: I6059ae35a77d675f54b98b2b43b5762e1d24365b
In [operation.set-admin-password], the parens in the following:
mechanism to re(set) the password
...imply it can be read as:
mechanism to re the password
...or
mechanism to reset the password
The former is obviously not as intended.
Change-Id: I696658a45d220a13a509d4f904d6b0fb996d7925
Cinder's volume migration API is, by default, an admin-only operation.
This includes the migrate_volume_completion API.
When Cinder is doing a volume migration, it calls Nova's swap-volume
API to detach the old volume that we're migrating from and attach
the volume that we're migrating to. Then Nova calls Cinder's
migrate_volume_completion API to signal Nova is done and Cinder
can finish the volume migration.
The problem is that swap-volume is not an admin-only API in Nova
per the default policy. So if a non-admin user tries to perform
a swap-volume operation, it will fail with a 403 when calling
Cinder's migrate_volume_completion API, since that requires an
admin user.
Also, because of 98739761f1 we can't
simply avoid calling migrate_volume_completion for non-migration
cases because that API handles the actual detach/attach for the old
and new volumes, swap-volume is broken without calling that.
So given swap-volume relies on an admin-only Cinder API, and is called
from an admin-only Cinder operation (volume migration), we should
just make it default to admin-only also.
Change-Id: Iac03258735f3d856a474ab96fe9b0a087e32906f
Closes-Bug: #1522705
Currently the rebuild action is supported in API and CLI of nova
though only ironic driver has its own implementation, we can still
list it into the support matix to provide more info to user.
Change-Id: I391b54a383501d3cdd2ac799d0ce4ccf66243f5a
Truely vmware driver can not attach cinder volume using chap
authentication over iscsi
Closes-Bug: #1557938
Change-Id: I05b1e81a3deffc855be34efff2d3e9dac8b63e82
How to trigger crash dump depends on hypervisors. NMI is not the only way to
implement the feature. This patch modifies description regarding this feature.
Change-Id: I77d10551650776c06ee4b413f1b027abf6620e83
The feature triggering crash dump has already been supported by nova [1],
now we should complete the support matrix correspondingly.
[1]The change id is: Iaf63352139c466430d3304a87a66398133541fac
Change-Id: I769cab81ba72d48db259d737cbe5e69f7d530e73
This patch update feature support matrix documentation for uefi boot.
Change-Id: I8ec6a2ad24e100cce48d70160350e5bf853992ef
Implements: blueprint boot-from-uefi
The set-admin-password virt driver API was implemented for libvirt
qemu/kvm x86 in liberty per 520a6b9285 but
the hypervisor support matrix wasn't updated.
This updates the support matrix and also adds notes that libvirt 1.2.16
is required along with hw_qemu_guest_agent configuration in the backing
image.
Closes-Bug: #1511114
Change-Id: Iee1d628116cad8c0c60c8335a2910b3c4e1b5e2d
Clarify in the hypervisor support matrix docs that "Stop instance CPUs"
is the same as "pause", and that "Resume instance CPUs" is the same as
"unpause".
Change-Id: I26d6a065dd239f376b57ae44d34276a994b30b2e
Closes-Bug: #1511487
It's not currently possible to create snapshots if you
use libvirt driver with parallels virt_type. First of all
it's because snapshot_extract method is not implemented
in imagebackend.Ploop class. There are also issues in
utils.extract_snapshot, find_disk and get_disk_type.
This patch solves all those issues, so snapshots will
work with parallels virt_type.
DocImpact: parallels now supports snapshot
Closes-bug: 1484074
Implements: blueprint ploop-snapshot-support
Change-Id: Id3b5bfc946a858f7e6ed710f2b44c7ca63a10723
As Parallels Cloud Server product was renamed within Parallels company rebranding
we need to change its reference in feature support matrix and documents accordingly.
A parameter 'virt_type' in libvirt section of nova.conf will remain 'parallels'.
Since libvirt version that supports 'vz' starts to report it as a hypervisor type,
we add a new hv_type 'vz' in the list of known hypervisors.
DocImpact.
Change-Id: I536e4284927508a813edd05f185f7ba0ed76d38c
implements: blueprint rename-pcs-to-virtuozzo
The libvirt Xen driver is partially supported, since it supports cold
snapshots, but it is said that the feature is missing. Changing it and
adding a note.
TrivialFix
Change-Id: Ie8b5c006129df4638f34b7c1fada5fc18ad672c7
The feature that was previously labeled as "evacuate" was in fact the
"host maintenance mode" which was triggered via a different API.
The real evacuate feature (with the evacuate API) is added by this
patch set.
http://lists.openstack.org/pipermail/openstack-dev/2015-June/067787.html
Change-Id: I525fa04f4b85cbe6af5b779c81d861fd56c0852f
The "serial console" feature was introduced during the Juno cycle
and is an alternative to remote access via VNC, RDP, SPICE for
platforms which don't support these.
This change adds a new feature row for that to the hypervisor support
matrix table. Because I'm not able to make a statement for all
hypervisor targets I use the "unknown" state for them.
Change-Id: I23381e30b6ddb823a50f7755179eb02bc18222e8
Some features could be made clearer if the CLI commands which are
related to that feature are listed. This patch set allows to set
CLI commands per feature.
This is an interim solution until the new feature classification
system is established.
Change-Id: If599627187080a8349811f50f6561e9efbc35138
Misled by the title of the feature "evacuate" the platform s390x
marked it as "complete". After a discussion in Nova IRC channel it was
clarified that this should be called "set host maintenance". This is
"missing" for s390x just like on x86. hyper-v doesn't implement the
necessary methods too.
http://lists.openstack.org/pipermail/openstack-dev/2015-June/067787.html
Change-Id: I3cd386269868c9fbbdbb16ca6281a86f4e2f1e0f
With commit da9301fbc the live-migration is possible for "libvirt KVM
(s390x)". This patch set updates the hypervisor support matrix for the
features "live-migration" and "evacuate" (which can use live-migration).
Change-Id: I17ef7012222c7088fa0ae750d06805e5254b1cf5
The document currently lists the ironic driver as partially supporting
the pause/unpause and suspend/resume operations, but those are not
available in the ironic driver in tree.
Change-Id: I6979b6e096b8320b817ddb71100b9f0162335be6
Wikipedia's list of common misspellings [1] has a machine-readable
version. This patch fixes those misspellings mentioned in the list
which don't have multiple right variants (as e.g. "accension", which can
be both "accession" and "ascension"), such misspellings are left
untouched. The list of changes was manually re-checked for false
positives.
[1] https://en.wikipedia.org/wiki/Wikipedia:Lists_of_common_misspellings/For_machines
Change-Id: I52a6f641a203185f9005bc0731d15166ad741161
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
We've added support for Parallels Cloud Server
recently, so let's add it to hypervisor support
matrix. Since it allows to work with both container
and hypervisor technologies and at this point lists
of supported features are different, let's add two
separate columns.
Change-Id: I6b4f7aa6490819a84d820d68149181123cd7b7c3
Two sections of the hypervisor support matrix refer readers to the
sections of the document dealing with the "pause" operation.
Although there is a section in the raw .ini file called
"[operation.pause]", this doesn't appear at all in the rendered
version of document since it's title is set to
"Stop instance CPUs". This renders readers of the rendered
document unable to find sections being referenced.
This patch clears up the confusion by changing the references
to refer to "Stop instance CPUs" rather than "pause".
Change-Id: Ia97d8e400db3a11291a715f6a630ea8bc638993e
Closes-bug: #1418641
Add document to replace / obsolete the giant table on
https://wiki.openstack.org/wiki/HypervisorSupportMatrix
This initial draft is a fairly straightforward conversion of
that table. Over time, it needs much work to improve the coverage
of API operations and and coverage of important configuration
information that users will care about.
It is using the .ini file syntax in order to record the data in
an easily machine parsable format, while remaining human friendly
by avoiding the syntax heavy approach of XML / JSON / YAML
An extension is registered with sphinx that can convert the
.ini file content into docutils content that then gets rendered
into the developer docs, linked from the index page
Change-Id: I4d3db4bce5737dba30a026a11083a9ea64459cd4
Jenkins