As a consequence of implementing multiple port bindings in the Neutron
reference implementation [1], the following attributes were made lists:
- 'port_binding' in the in the SQLAlchemy Port model.
- 'binding' in the Port OVO
As a consequence of this change, these attributes names were changed to
'port_bindings' and 'bindings' respectively. This patch updates the
references to those attributes.
To support neutron envs, both with and without the neutron change [2],
this patch adds a shim to use the proper port binding accessor.
NB: this shim may allow the tests to pass but does not test using the
proper "depends on" approach since that doesn't work today with the 3rd
party VMware NSX CI.
For more details see [1][2].
[1] https://review.openstack.org/#/c/571041
[2] https://review.openstack.org/#/c/414251
Change-Id: I35c24b83150d84a15997c5ebe3e7b51015f87e99
This patch updates tox ini to support both a dev and build target;
lower-constraints-dev and lower-constraints respectively.
The dev target is for running locally so it manually installs the
required projects whereas the build target doesn't as per zuul's
required project definition in project.yaml.
Also see https://review.openstack.org/#/c/573386/
Change-Id: I0e6cf1ca0f4b07c080ff5a7b221dabedfd8bf6cf
This patch updates our local zuul.yaml to define the required projects
so they are pulled into the jobs.
Change-Id: I1e668f9413363954f148f35f88c4c0687ce4f2c0
The recent Octavia preparation caused the LB drivers not to have
a core_plugin property which fails the on-router-action callbacks.
This patch will address it.
Change-Id: I6b2c98bd605ff33bf173911898267a8ed3f7045f
Currently the plugin does not support shared routers with FWaaS firewalls.
So it should fail when the user tries to change the type of a
router with a firewall to 'shared'.
Depends-on: Ib31210121a56561fb33a8ed0a1fbbdf6b01db727
Change-Id: Ieeed46ffd7309a8a8d423e44f766fec5f47b6184
Commit I4ebd24f1b13eb823c4d63452fd37cace5bcf5481 moved some files
that the FWaaS callbacks import.
Change-Id: Iec0d68f2d62f1f32e12e09db40253f5545f8fcec
When trying to add interface to another distributed
router, a proper message should notify the user
that this is incorrect and not supported.
Change-Id: If9230c821665813939e676fbaafe296d42cac498
Signed-off-by: Michal Kelner Mishali <mkelnermishal@vmware.com>
Add a lock around operations that read LB pools from the backend
and adjust members to add/update/remove items.
This ensures that in case of concurrent operations no stale items
could be added back to the pool accidentally.
Change-Id: I91b5c1d9eb4493c0d19dadaca0af271d0a2afe5c
When a router is deleted, we check if any of its subnets is used
by LBaaS.
Since normally there are no subnets at this point, the check fails.
Should verify the subnets existence in teh callback.
Change-Id: I2c23722e616a8a05f6fc6d24336356355b96ca9b
This patch bumps the neutron-lib required version to 1.16.0, mainly to
pick up the fix in https://review.openstack.org/#/c/574068/
Change-Id: I8c445f7214164ff91c1513cb019d37b167eb0c9f
Create internal drivers for each object that will use dictionaries
instead of objects, and will use a generic fail/success notifiers.
The LBaaS nsx-v/v3 drivers will translate the objects to dictionaries
and use the LBaaS callbacks for successful_completion and failed_completion.
Later we can add drivers for Octavia that will use the same internal
drivers.
Change-Id: I5f99f979bd5333ccfc781217379d217d611cb015
Commit I0775e284f21012a01298168d5ca3dd1c74b95f05 added an import of
the mock package, so it should be added to the requirements
Change-Id: I444ab50d5bdb20c528c3a1581dc7e9301ab06dad
The vmware-nsx project is the only one using safe_reference [1] from the
common db mixin. We plan to remove safe_reference from the mixin, so
this patch switches vmware-nsx over to use weakref.proxy directly.
[1] http://codesearch.openstack.org/?q=%5C.safe_reference
Change-Id: I921c9265a646c3c291813d4488ad0fe0143038ca
When initialzing the default NSX section and NS group, the plugin
create a default global security group, in order to make sure no
other server is doing it at the same time.
In case anotehr server does, we get a DP duplicated entry error,
but after retry it somehow works, causing the 2 process to handle
the backend at the same time.
This patch avoids the retry in case of DB errors for this case.
Change-Id: I0775e284f21012a01298168d5ca3dd1c74b95f05
When creating a net and subnet on py35 env,
an error occurs noting that str should be
bytearray
Change-Id: I069a24f5a5346ce3cee9903b1e5a4fd16ffa9f86
Signed-off-by: Michal Kelner Mishali <mkelnermishal@vmware.com>
When retrieving NNSX backend objects, we should assume that not
all of those have the tags property
Change-Id: I6c3bc8d8735604ba463e57d7dbd9409028d30ef1
Since the FWaaS v2 pluggable driver patch
I4ebd24f1b13eb823c4d63452fd37cace5bcf5481
The initalization of the dummy FWaaS callbacks for admin utilities
should change.
Change-Id: I4d7bf6e470f38ea6d677fcb708ca0117b7539fa2
Backend doesn't produce NotFound exception when deleting a virtual server
or an application profile which doesn't exist.
Instead it produces RequestBad exception which should be caught and handled.
Change-Id: I9f3e7e55ddfea470c69c46b95dd0ac64136ecf97
