1. What is the problem
In the patch[1] for solving bug[2], we introduce an error that when
a fip is created with port ID and the bottom port is not there, the
fip body is not returned. Our smoke test doesn't find out this error
because it only affects the response, the fip is correctly created
in the database. But the log does show an error message.
2. What is the solution for the problem
Always return fip body. Unit test is modified to use the returned
body to ensure it's not None.
3. What features need to be implemented to the Tricircle to
realize the solution
N/A
[1] https://review.openstack.org/#/c/465870
[2] https://launchpad.net/bugs/1691918
Change-Id: I71fcc8fab9fb2a50ab058d67980f6e1e1e4e6d90
1. What is the problem
The QoS service will release in Pike, and the spec now in Ocata folder.
2. What is the solution to the problem
Move QoS spec from Ocata folder to Pike folder.
3. What the features need to be implemented to the Tricircle
to realize the solution?
QoS.
Change-Id: I347a3a5460541f4ca69cdca8b8c2633bbcd3a4b4
1. What is the problem
(1) In this patch[1] of Neutron project, new object model for security
group is used. As a result, the "remote_ip_prefix" field returned
by "get_security_group_rule" is an AuthenticIPNetwork object instead
of a string, so our "_compare_rule" method fails to recognize two
identical rules.
(2) After this patch[2] of devstack-gate project, horizon is no longer
the default project to be deployed in gate test, so enable horizon in
our own local.conf without explicitly enabling horizon in the gate test
will cause gate test to fail.
2. What is the solution for the problem
(1) Use str() to transform the AuthenticIPNetwork object to string.
(2) Disable horizon in our own local.conf since we don't need it.
3. What features need to be implemented to the Tricircle to
realize the solution
N/A
[1] https://github.com/openstack/neutron/commit/
af52d499a53f9dddacd8c9116d1bb0570e8f579c
[2] https://github.com/openstack-infra/devstack-gate/commit/
e1b6e9f1d6532fe87a8947a633e254d14aeb8fc0
Change-Id: I57ac130cb2383f2912fc6c0af646fb84ab2cdd90
1. What is the problem
QoS is defined as the ability to guarantee certain network requirements
like bandwidth, latency, jitter, and reliability in order to satisfy a
Service Level Agreement (SLA) between an application provider and end
tenants. Networks or ports in local Neutron can not be associated with
QoS policy automatically yet.
2. What is the solution to the problem
Using asynchronous jobs to complete the automation of QoS policy/rule
in local Neutrons.
3. What the features need to be implemented to the Tricircle
to realize the solution?
QoS.
Change-Id: I8dcce7d35d7fe7283cdeb80d71259d3f3859406b
1. What is the problem
In smoke test, we use database query to check if there are running
jobs. OpenStack CI may use different database backends so we need
to write scripts for different backends.
2. What is the solution for the problem
Job API has been merged, we can use it to check job status.
3. What features need to be implemented to the Tricircle to
realize the solution
No new features
Change-Id: I3eeb05f3fb4ea188eea4633b64f0f22332887559
1. What is the problem
After booting a VM, we associate a FIP to it via central Neutron.
However, local FIP is not created as expected. Also, there is no related
asynchronous job registered.
The reason may be that we don't trigger asynchronous job in
"create_floatingip" method, and since VM is already there, there's no
chance to trigger the job.
2. What is the solution for the problem
If "port_id" is specified in the body, trigger asynchronous job.
3. What features need to be implemented to the Tricircle to
realize the solution
Support creating FIP with "port_id" specified.
Change-Id: Iceb15f68acc23f8dcb8767ac0371947334a1e9db
Closes-Bug: #1691918
1. What is the problem
In our unit tests, test_central_trunk_plugin and test_central_plugin
share many codes in FakeSession, FakeQuery and FakeClient.
2. What is the solution for the problem
Move FakeSession, FakeQuery and FakeClient to a common module so we
can reuse them in the later development of unit tests.
3. What features need to be implemented to the Tricircle to
realize the solution
N/A
Change-Id: I79078cc5a561f95b1e46a046c0bf0babf9d38363
DeprecationWarning:
Using function/method 'oslo_messaging.transport.get_transport()'
is deprecated: use get_rpc_transport or get_notification_transport.
Replace get_transport with get_rpc_transport.
Change-Id: I18eddf2cb1ae7f50f2c5c9ccf30b87cadc34d3a3
Following OpenStack Style Guidelines:
[1] http://docs.openstack.org/developer/hacking/#unit-tests-and-assertraises
[H203] Unit test assertions tend to give better messages for more specific
assertions. As a result, assertIsNone(...) is preferred over
assertEqual(None, ...) and assertIs(..., None)
Change-Id: I378da84574fd06d873c72bb9b05cfe7da75c4cfe
* The tricircle projects contains the skeletion of tempest plugin and
there is nothing implemented with in the plugin and the existing plugin
extension is breaking the tempest cli. So it is better to completely
remove the plugin.
* In future, if we need tempest plugin for tricircle, we need to implement
in a seperate repo.
Change-Id: I2dbf532ae9b33035187f92b9092af6d26eaba338
Closes-Bug: #1693546
1. What is the problem
We try to figure out a way to integrate Tricircle with Nova cell v2.
2. What is the solution for the problem
The basic idea is to start a local Neutron server for each cell. Nova-
compute in each cell is configured to talk to local Neutron server in
the same cell. All local Neutron servers are configured to talk to the
very one Nova-API.
Currently DevStack doesn't support multi-cell deployment, so we try to
deploy it with our own plugin. In node1, Nova services are started as
before, but we change the region of Nova-API, Glance-API and placement-API
from RegionOne to CentralRegion. Local Neutron server is also configured
to talk to Nova-API in CentralRegion. In node2, Nova-API is enabled at
first, since we need DevStack to help us create the related database.
After the DevStack start in the node2, we manually stop Nova-API and
Nova-scheduler in node2.
One document discussing the detailed setup and trouble shooting steps
is added.
3. What features need to be implemented to the Tricircle to
realize the solution
Tricircle can work with Nova cell v2.
Change-Id: I6ba8e1022d83f40df36464abfdd7b4844673b24d
1. What is the problem
Set external gateway to a router, if error occurs during local router
creating/updating, exception will be raised, but this time central
router is already updated with external gateway information. So when we
try to unset the external gateway, central Neutron server will try to
remove external gateway for local router first, if local router doesn't
exist at this time, we have no way to unset the external gateway since
we fail to remove external gateway for local router.
2. What is the solution for the problem
Do not try to remove external gateway for a nonexistent local router.
3. What features need to be implemented to the Tricircle to
realize the solution
N/A
Change-Id: Id7261640405d4d24c8523f6518ec19edc031024b
Closes-Bug: #1693138
1. What is the problem
We can create a subnet without gateway in the central Neutron, but when
we boot a VM in that subnet, the request fails since local Neutron is
trying to calculate the IP allocation pool based on the central subnet
gateway IP, which is None.
2. What is the solution for the problem
If a central subnet doesn't have gateway, do not re-calculate local IP
allocation pool when creating local subnet.
3. What features need to be implemented to the Tricircle to
realize the solution
Support booting a VM in a subnet that has not gateway
Change-Id: I04f7f9f9926f22bca3490ee46e2163b9c17bb8fe
Closes-Bug: #1693446
1. What is the problem?
There are no unified terms used in Tricircle networking, this
will lead to confusion for the end user.
2. What is the solution to the problem?
Add networking terms to ease Tricircle networking automation,
reduce the misunderstanding on the networking guide.
3. What features need to be implemented to the Tricircle to realize
the solution?
None
Change-Id: Ie857d7ce6cab7ab4545da6363cc70eb8c621d755
Signed-off-by: joehuang <joehuang@huawei.com>
Probably the most common format for documenting arguments is reST field
lists [1]. This change updates some docstrings to comply with the field
lists syntax.
[1] http://sphinx-doc.org/domains.html#info-field-lists
Change-Id: I6b783d7c92852d656586e30d2456218436d1a015
[1] moves the attr decorator from test.py to tempest/lib. So, all the
references to tempest.test has to be moved to tempest.lib.decorator.
[2] https://review.openstack.org/#/c/456236/
Change-Id: I26a45e54dbb58bd975a403eb62fdba0dde165fa6
1. What is the problem?
The python-tricircleclient has been included as part of the OpenStack
official repositories[1] but their integration hasn't been done in
Devstack.
[1] https://review.openstack.org/#/c/426419/
2. What is the solution to the problem?
Add the required instructions to control the installation of the
client in the plugin.sh script. Usually those instructions have a
standard format which needs to be follow.
3. What the features need to be implemented to the Tricircle
to realize the solution?
None
Change-Id: I0f2df46316f191c443b9192875868d4c7be96c1b
1. What is the problem?
Provider attributes have been already move to Neutron-lib,
but we still have some hard-coded string: 'provider:network_type',
'provider:physical_network' and 'provider:segmentation_id'.
2. What is the solution to the problem?
We can use the constants in neutron_lib instead
3. What features need to be implemented to the Tricircle to realize
the solution?
None
Closes-Bug: #1691634
Change-Id: I4da0064d2ac78224698e62b8bd42935c6f8f25c9
1. What is the problem?
There are so many if branches for az_name in pod.py,
this will lead to confusion.
2. What is the solution to the problem?
We can optimize them to keep the logic more clearly
3. What features need to be implemented to the Tricircle to realize
the solution?
None
Closes-Bug: #1691653
Change-Id: I1f31ff685de2e516d36027388a7b0eacc1a76f5e
1. What is the problem
In resource_handle.NeutronResourceHandle, after connection-fail
exception is caught, endpoint_url attribute is set to None and waits
for client.Client to update the endpoint.
However, local plugin directly uses resource_handle to talk to central
Neutron, so the endpoint_url attribute will not be updated after it's
set to None. Actually, for local plugin, endpoint_url should be read
from the configuration file and doesn't require updating, so it's not
necessary to set it to None.
2. What is the solution for the problem
Reset endpoint_url in Client instead of ResourceHandle
3. What features need to be implemented to the Tricircle to
realize the solution
N/A
Change-Id: If306031605f30adf933464b955386162c7f83c06
Closes-Bug: #1691306
1. What is the problem
For resource routing Admin API, when we call get_all(self, **kwargs)
function with an unsupported filter type specified to list the routing
information, the function will ignore this filter and return all the
routing information. It shouldn't happen. Instead, when unsupported
filter type is given, the function should response with 400 error
and notify user there exist unsupported filter types.
2. What is the solution to the problem
Response with 400 error when unsupported filter type is specified for
routing's get_all(self, **kwargs) function.
3. What the features need to be implemented to the Tricircle
to realize the solution
Modify the original resource routing's get_all(self, **kwargs)
function to return 400 error when met unsupported filter type.
Change-Id: I1bff9ebb7a6371a139d9ba547af417548100a2b4
Closes-Bug: #1690971
1. What is the problem
pod_id in ShadowAgent is a foreign key, it references pod_id in Pod but
has different string length with it.
2. What is the solution to the problem
Change pod_id string length to 36 in ShadowAgent to keep consistent
with its definition in Pod.
3. What the features need to be implemented to the Tricircle
to realize the solution
None.
Change-Id: I75c7f0680e9c451a3df31d8a174007bf0a193e0f
Closes-Bug: #1691366
1. What is the problem
As reported in the bug page, local subnet will be recreated
after central subnet is deleted
2. What is the solution for the problem
The reason for the recreation is that DhcpAgentNotifyAPI subscribe
the subnet deletion event and will call get_network during subnet
deletion, get_network will call _ensure_subnet to create the subnet.
So we need a way to distinguish the call from API and callback.
A dict on_subnet_delete is added to local plugin in this patch.
In delete_subnet, before calling real core pluing's delete_subnet
method, we first add a new item (request_id, True) to this dict.
After the calling, this item is removed. So in get_network, we can
check whether the dict contains the request_id, if so, we know
that we are in the subnet deletion stage, then we don't call
_ensure_subnet.
3. What features need to be implemented to the Tricircle to
realize the solution
No new features
Change-Id: Ie79460aa763918be7c522263d9865847fc0da828
Closes-Bug: #1680757
1. What is the problem
If a tenant creates two subnets with the same CIDR but different
allocation pools, Xjob will fail to setup local security group rules
because it generates two identical rules for these two subnets but
Neutron disallows adding a rule which is identical to the existing one.
2. What is the solution for the problem
Use "set" to avoid generating duplicated rules
3. What features need to be implemented to the Tricircle to
realize the solution
N/A
Change-Id: If21562458f2ad82e71577392172deeb8dc5a2a13
Closes-Bug: #1691315
