Sometimes it can happen that when extending a volume it is shown as
status "extending" with the original size for a while. Wait for the
volume to enter state "available" again before checking the result.
Signed-off-by: Dr. Jens Harbott <harbott@osism.tech>
Change-Id: Ib70cfa1c241ce94426214c7a05c550213f427caa
The old novaclient.v2.server.Server.interface_attach() method is
replaced with proxy.create_server_interface().
In swargs, 'net_id' and 'port_id' are mutual-exclusive, if one of
them is given with value, the other one cannot be None, as the
API would responde with 400 (None is not string).
In unit test, temporary method 'setup_sdk_servers_mock' is added,
because other tests are still using the old 'setup_servers_mock'.
Functional tests are added. Releasenote is generated.
Change-Id: I9899f0509febc5143560a1859ae6344d0a6d1427
Cinder recently removed their v2 API [1] which is causing the functional
tests to fail. Improve our 'is_service_enabled' test helper to use the
'versions show' command, which queries the service catalog and can give
us information about the service version as well as answer the more
general "is this service available" question. We also resolve a
long-standing TODO in the process.
[1] https://review.opendev.org/c/openstack/cinder/+/792299
Change-Id: I381069357aa008344e15327adf3a863c0c2e1f04
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Neutron has got CRUD API for L3 conntrack helper since some time.
This patch adds support for it in the OSC.
OpenStack SDK supports that since [1]
This patch also bumps minimum OpenStack SDK version to
the 0.56.0 as that version introduced support for the
Neutron's L3 conntrack helper.
[1] https://review.opendev.org/c/openstack/openstacksdk/+/782870
Change-Id: I55604182ae50b6ad70c8bc1f7efad8859f191269
Devstack recently switched default Neutron's backend from ML2/OVS to
ML2/OVN. As OVN backend has some parity gaps and differences in some
APIs, functional tests job was failing with ML2/OVN as some tests
weren't properly skipped in case of missing some Neutron API extensions.
This patch fixes that by doing some small changes in the functional
tests:
- skip DHCP/L3 agent tests when dhcp/l3 agent scheduler extensions
aren't available,
- skip updating neutron agent as OVN agents don't allows that,
- skip service providers tests when there is no Neutron L3 agent
available,
- skip setting router as distributed as OVN backend don't supports that
router's attribute at all.
Depends-On: https://review.opendev.org/c/openstack/neutron/+/793141
Change-Id: I29a8db202086b0b49fed865409fa8ca244b98439
This mostly reuses the existing tests for '--block-device-mapping',
which can hopefully be removed at some point in the future.
This highlights two issues with the implementation of this option.
Firstly, the 'boot_index' parameter is not required so don't mandate it.
Secondly, and more significantly, we were defaulting the destination
type for the 'image' source type to 'local'. Nova only allows you to
attach a single image to local mapping [1], which means this default
would only make sense if you were expecting users to use the
'--block-device' option exclusively and omit the '--image' option. This
is the *less common* case so this is a bad default. Default instead to a
destination type of 'volume' like everything else, and require users
specifying '--block-device' alone to pass 'destination_type=local'
explicitly.
[1] https://github.com/openstack/nova/blob/c8a6f8d2e/nova/block_device.py#L193-L206
Change-Id: I1718be965f57c3bbdb8a14f3cfac967dd4c55b4d
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
The '--offload' option allows us to explicitly request that the server
be offloaded once shelved or if already shelved.
The '--wait' option allows us to wait for the shelve and/or offload
operations to complete before returning. It is implied when attempting
to offload a server than is not yet shelved.
Change-Id: Id226831e3c09bc95c34b222151b27391a844b073
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Make use of 'FormattableColumn'-derived formatters, which provide better
output than what we were using before, particularly for the YAML output
format.
Change-Id: Id6d25a0a348596d5a0430ff7afbf87b049a76bc8
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Make use of 'FormattableColumn'-derived formatters, which provide better
output than what we were using before, particularly for the YAML output
format. For example, compare before for the 'server show' command:
$ openstack --os-compute-api-version 2.79 server show test-server -f yaml
...
addresses: private=fdff:77e3:9bb4:0:f816:3eff:fe6d:a944, 10.0.0.44
flavor: disk='1', ephemeral='0', extra_specs.hw_rng:allowed='True', original_name='m1.tiny',
ram='512', swap='0', vcpus='1'
...
To after:
$ openstack --os-compute-api-version 2.79 server show test-server -f yaml
...
addresses:
private:
- fdff:77e3:9bb4:0:f816:3eff:fe6d:a944
- 10.0.0.44
flavor:
disk: 1
ephemeral: 0
extra_specs:
hw_rng:allowed: 'True'
original_name: m1.tiny
ram: 512
swap: 0
vcpus: 1
...
Similarly, compare before for 'server list':
$ openstack --os-compute-api-version 2.79 server list -f yaml
- ...
Networks: private=fdff:77e3:9bb4:0:f816:3eff:fe6d:a944, 10.0.0.44
Power State: Running
Properties: ''
...
To after:
$ openstack --os-compute-api-version 2.79 server list -f yaml
- ...
Networks:
private:
- fdff:77e3:9bb4:0:f816:3eff:fe6d:a944
- 10.0.0.44
Power State: 1
Properties: {}
...
We also fix the human-readable output for the 'tags' field.
Before:
$ openstack --os-compute-api-version 2.79 server list
...
| tags | ['bar', 'foo'] |
After:
$ openstack --os-compute-api-version 2.79 server list
...
| tags | bar, foo |
Change-Id: I7a8349106e211c57c4577b75326b39b88bd9ac1e
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Add support for Neutron address group CRUD operations. Subsequent patches
will be added to use address groups in security group rules.
Change-Id: I3c313fc9329837dde67815901528a34dca98ebcc
Implements: blueprint address-groups-in-sg-rules
Depends-On: https://review.opendev.org/738274
Depends-On: https://review.opendev.org/745594
Currently, we attempt to unset an image property by setting it to None.
This doesn't work for known properties and is rightly rejected by the
Glance API with the following error:
BadRequestException: 400: Client Error for url:
http://172.20.4.87/image/v2/images/368c5751-2b0b-4a38-a255-fd146fe52d31,
Bad Request
The solution is to actually unset the field by deleting it.
Change-Id: Ie156bedbe0f9244f82c81401679706f484caf9aa
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Story: #2008463
Task: #41493
The 'os-agents' API was recently removed from nova [1]. Remove the
functional tests, since they will always fail going forward but will
continue to run on older stable branches.
Also Squeeze https://review.opendev.org/#/c/762559/ inside, since those
2 are simultaneously blocking gate
[1] https://review.opendev.org/#/c/749309/
Change-Id: I0bf7d4c0ba2a9d4637db0d209d8d52163d772f12
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Change I1d1ac1ac46f49f64794ffc8631e166935537966c introduced the 'quota
set --force' parameter to force set nova quotas. As part of that fix, we
introduced a functional test, 'QuotaTests.test_quota_set_force' that
works by attempting to set the 'limit' of the quota for instances to the
current usage ('is_use') minus one.
This test is flawed. It doesn't create any instances so when it fires by
itself, it will always set the 'limit' to 0. When it fires at the same
time as other tests (remember, we run tests in parallel), notably tests
that rely on booting instances, it can cause other tests to fail with
the following error:
Quota exceeded for instances: Requested 1, but already used 0 of 0
instances (HTTP 403)
We could attempt to work around this by creating a new project and using
that project to fiddle with quotas. That's a lot of work though, and the
returns are questionable: the 'quota set' command is an admin-only
command by default and the '--force' parameter should almost never be
used. Simply remove this test.
Change-Id: Ic07ff6f4a7c1c27852c892eb906bb144aae91788
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Story: #2008327
Task: #41225
Fixes incorrect json output for 'openstack server show -f json'.
The security group json output groups all the json as one
for e.g. "security_groups": "name='group1'\nname='group2'"
The correct output should be
"security_groups" : [{"name" : "group1"}, {"name" : "group2"}]
properties and volumes_attached fields also has similar issue.
Story: 2007755
Change-Id: I1b1cac716329e0530400aff782c08000b21d8e1d
For a server booted from a volume, nova API does not store an image_id
and instead returns an empty string. Currently, openstackclient
similarly shows an empty string for Image Name and Image ID for servers
booted from volumes.
To aid CLI users in understanding the meaning of no image_id, we can
display the string "N/A (booted from volume)" in the image field if the
server was booted from a volume.
Change-Id: I9c62cf6fe23b2e934dcbf5ebbf706b2705d2e424
The file test_examples.py has never worked since its written and and cli in
the example directory are covered by other functional tests for container,
flavor and object lists and they have better asserts. So, deleting the file
Change-Id: Ib9af40da96e66354fe878e79a80048a58f8dd6fe
No idea how this happened, but reviews started failing the
pep8 gate job. The failures are legitimate, see the commit.
I guess the pep8 tests became smarter and found these issues.
Change-Id: Id9a0dad644134dafd68eed37fe8f41c583d7a619
The compute service allows us to to force set a quota, setting a
quota value that is less than the amount of the resource currently
consumed. Expose this feature by way of a '--force' boolean
parameter.
Change-Id: I1d1ac1ac46f49f64794ffc8631e166935537966c
Add support for stateful attribute of security groups,
using --stateful and --no-stateful flag on security group.
This allows a user to create security groups with stateful
false.
Change-Id: Ifd20b5fc47fd0ea0bb5aeda84820dcc0fb1e8847
Blueprint: stateless-security-groups
Depends-On: https://review.opendev.org/711513/
Add commands to osc to call the two new API methods introduced by
new Neutron extension: extraroute-atomic.
Bump our openstacksdk requirement to >=0.38.0 which contains
the corresponding sdk change.
The lower-constraints of dogpile.cache and keystoneauth1 are bumped
because of requirements bumps in openstacksdk.
The lower-constraint of decorator is bumped because of problem already
fixed by amotoki here: https://review.opendev.org/701706
Change-Id: Ia9b9c216f1d1161ebedac31594a2c464d77f4ae2
Depends-On: https://review.opendev.org/674324
Partial-Bug: #1826396 (rfe)
Related-Change: https://review.opendev.org/655680 (spec)
This patch removes using the "name" option for a marker when
--deleted is also used. The find_resource() function
that is being called does not correctly handle using the marker
as the "name" in the search when also using deleted=True.
One simple way to fix this is force the marker to only be an ID
when --deleted is used. This is how the nova client works.
Using the --deleted option is available to users with the admin
role by default. If you're an admin listing --deleted servers
with a marker by name, find_resource() is going to fail to find
it since it doesn't apply the --deleted filter to find_resource().
The find_resource() function is trying to find the marker server
by name if it's not found by id, and to find it by name it's
listing servers with the given marker as the name, but not
applying the --deleted filter so it doesn't get back any results.
In the story it was suggested modifying find_resource to include
the deleted query param when it's specified on the command line but
that didn't work because it still results in something like this:
http://192.168.1.123/compute/v2.1/servers?deleted=True&name=4cecd49f-bc25-4a7e-826e-4aea6f9267d9
It seems like there are bugs in find_resource().
Restricting the marker to be the server ID when listing deleted servers
is probably OK since if you're using --deleted you're an admin and you could
be listing across all projects and if you're filtering by a server across all
projects anyway (not that you have to, I'm just saying if you are), or even
showing a server in another project, you have to do it by id rather than name
because find_resource() won't find the server in another project by name, only ID.
story: 2006761
Task: 37258
Change-Id: Ib878982b1d469212ca3483dcfaf407a8e1d2b417
The correct state name for a failing volume snapshot deletion is
'error_deleting' instead of 'error-deleting'. [1]
[1] 89d6a5042f/cinder/objects/fields.py (L126)
Task: #37844
Story: #2007037
Change-Id: Ia99900ece4f1cd29769b22ddaa3965789d719556
Until now-ish Volume v3 has been a pass-through to v2. In order
to prepare to make the Volume v3 commands stand-alone copy the
v2 functional tests to v3.
This is the first of a series of reviews to completely separate
Volume v2 and v3 commands. Once these are split we can begin to
implement v3 microversion support and/or start using the
OpenStack SDK as the REST library.
Change-Id: Iefd78d8ef6bb851d7360596337a88ee8f8476767
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
While investigating the referenced story/bug I noticed that
wait_for_status in
openstackclient.tests.functional.compute.v2.test_aggregate.AggregateTests
was doing a lot more than it should ever need to (it probably got copied
in from somewhere). The two places calling it only need to a) check the
output of `openstack aggregate show`, and b) try once -- since they just
got done creating the aggregate synchronously, there should never be a
need to delay/retry. So this commit removes the helper method and just
inlines the check.
At the same time, the addCleanup(aggregate delete) directives are moved
above their respective creates. This is a defensive best practice which
makes sure cleanup happens even if something fails very soon after the
actual back-end create (as was in fact the case with the referenced
bug/story).
It is unknown whether this will impact the referenced bug.
Change-Id: I0d7432f13642fbccd5ca79da9c76adfcbabb5fa9
Story: 2006811
Related-Bug: #1851391
Fix various things so the functional tests will work under python3:
- A hashlib.md5() can only be update()d with an encoded string in py3.
- There's no dict.iteritems(), change to dict.items() (which is already
an iterator).
- Open temp files with 'w+' mode rather than the default 'w+b' (as an
alternative to encoding all the write and expected-read payloads as
bytes).
- (This is a weird one) Explicitly raise SkipTest from unittest (rather
than unittest2, which is where cls.skipException landed). Not sure why
this is busted, but this moves the ball.
Change-Id: Ic9b2b47848a600e87a3674289ae7ae8c3e091fee
AggregateTests.wait_for_status() was a classmethod, those often
are sources of conflict in parallel testing...
Change-Id: I6211fd9c36926ca97de51a11923933d4d9d2dfda
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
Multiple subnets with the same name are occasionally created when
running tests in parallel.
Change-Id: Ifb85e39ee53b529e2b97abf782c7fba93d48e9e2
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
The token_endpoint was a compatibility auth type to maintain support
for the --url global option that dated back to the beginning of
OpenStack CLI auth. The common keystoneauth library implements
'admin_token' which provides the same functionality using
--endpoint rather than --url.
Change-Id: I1b9fbb96e447889a41b705324725a2ffc8ecfd9f
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
Rename metadata to property in all aggregate commands
Beef up functional tests to reduce street racing
Change-Id: I4598da73b85a954f3e6a3981db21891b45d9548c
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
Add these tests before hacking on the global args and removing
the compatibility stuff so we can clearly see what actually changes.
Change-Id: Ic86c89da1475b4914ff7cb2396199cd219a12097
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
Change the functional test TestCase.openstack() method to add a
way to not include the --os-auth-type option in order to test the
default auth-type logic.
Change-Id: I0f1ca2f7517a41278afaad5aaf4e98accb16bea2
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
This adds a --boot-from-volume option to the server create
command which is used with the --image or --image-property
option and will create a volume-backed server from the
specified image with the specified size. Similar to the
--volume option, the created root volume will not be deleted
when the server is deleted. The --boot-from-volume option
is not allowed with the --volume option since they both create
a block device mapping with boot_index=0.
Change-Id: I88c590361cb232c1df7b5bb010dcea307080d34c
Story: 2006302
Task: 36017
