integrate with neutron patch Id3f09b78c8d0a8daa7ec4fa6f5bf79f7d5ab8f8b
And also skip new tests added in I99681736d05eefd82bdba72b3866eab9468ef5dd
Change-Id: I8b119bc69cc87185ea77646e70135c5984200038
1.Upgrade pylint to 2.4.4, add exclusions to the tests, and
fix some lint errors in the code
2. Fix user creation with GRANT in MySQL 8.0(Ubuntu Focal)
In Ubuntu Bionic (18.04) mysql 5.7 version used to create
the user implicitly when using using the GRANT.
Ubuntu Focal (20.04) has mysql 8.0 and with mysql 8.0 there
is no implicit user creation with GRANT. We need to
create the user first before using GRANT command.
See also commit I97b0dcbb88c6ef7c22e3c55970211bed792bbd0d
3. Remove fwaas from the zuul.yaml
4. Remove DB migration test which is failing ue to FWaaS migration
with py38
5. Fix cover tests python version in .tox
6. fix requirememnts
Change-Id: I22654a5d5ccaad3185ae3365a90afba1ce870695
Since py2 is no longer supported, built in methods can replace the
six package usage, as been done in the neutron project
Change-Id: I922963fbbcc0ab263e1f6e56907b73b007015a75
For the fire cell anti affinity to work as designed, there is a need to use different
groups & rules per host group, since those hostgroups can be different for differnet
availability zones
Change-Id: I092f5c228489a3a0d73f060380f1a1a6c526fb00
create_dhcp_binding may fail with timeout or other types of exceptions
which do not have a response field.
This patch checks its existince before trying to decode it.
Change-Id: Ie62d3965a10bf07795c1bd56b7c6191eb5c73251
A recent change in pep/pycodingchecks introduced new warnings as part of
the pep8 target that causes pep8 to fail now.
This patch fixes code that issued warnings W503,E731,E266,E402
Change-Id: I57c035440bd847193ea5f8a8078016fb9baa3c31
When there is a failure during the recycling of an edge appliace to the
backup pool, the edge at the backend may still be attached to networks
and use IP addresses which are free for reuse by Neutron.
Housekeeping job should address such cases.
Change-Id: I3a8ba622f742064bdc8906ba745da0a54a4576ac
Adding a new configuration parameter to define the appliance size used for
new shared routers: shared_router_appliance_size
Change-Id: Ie28877469439917bf6df91702fd6608c78562d0d
The nsx-v backend does not suport adding a route where the
nexthop is part of the destination.
This patch skips it (instead of failing at the backend)
and also skips a tempest test that tries the same thing
Change-Id: I41181d43cffbf0e5cbb63069ed864683834c77b9
Commit I81748aa0e48b1275df3e1ea41b1d36a117d0097d added the l3 extension
API definition to neutron-lib, commit
I2324a3a02789c798248cab41c278a2d9981d24be rehomed the l3 exceptions and
Ifd79eb1a92853e49bd4ef028e7a7bd89811c6957 shims the l3 exceptions in
neutron.
This patch consumes the l3 API definition from neutron-lib
in prep for If2e66e06b83e15ee2851ea2bc3b64ad366e675dd
Change-Id: I56409beb595c69f2ecaac5349dea35bee14a26ce
When deleting a network, the plugin delets the rotuer binding entry
before deleting the dhcp port.
So we should not log a warning about not finding the dhcp edge.
Change-Id: I7a58a6789d59055ce1782c987bd359756b8bc16b
When something fails during the init_complete process, the plugins
md_proxy data is not fully initialized, so it is possible that even the default
handler was not set yet.
This patch ensures that the relevant md-proxy handler exists before using it.
Change-Id: I1db84c0abc30d8ea3d601f26b5b852a254a6036c
Ensure that the vnic incides are lock when removing the network.
This may cause a race when multiple networks are deleted at the
same time.
Change-Id: I13c01947ebe2762c76a75cdebb023aa7fed88aca
If an edge is stuck in PENDING status for too long, we are deleting it.
In case of PENDING_UPDATE or PENDING_DELETE we should compare the current time
the updated_at field and not created_at
Change-Id: Iec02c254233fd55362a8c51b6c153e7b6a2ee083
NSX 6.2.8 and 6.3.3 support the granular API:
New GET api:
GET https://<nsxmanagerIp>/api/4.0/edges/<edge-id>/dhcp/config/bindings/<binding-id>
RESPONSE:
<staticBinding>
<autoConfigureDNS>false</autoConfigureDNS>
<leaseTime>86400</leaseTime>
<subnetMask>255.255.255.0</subnetMask>
<bindingId>binding-1</bindingId>
<vmId>vm-34</vmId>
<vnicId>1</vnicId>
<hostname>test</hostname>
<vmName>3-vm_ubuntu_1404_srv_64-shared-1668-02458540-04f1-4508-8037-2f2ce5542b91</vmName>
<ipAddress>12.12.12.1</ipAddress>
</staticBinding>
2) Quering a binding with a invalid ID:
GET https://<nsxmanagerIp>/api/4.0/edges/<edge-id>/dhcp/config/bindings/<binding-id>
RESPONSE:
<error>
<details>[Dhcp] Static binding is not found: binding-2</details>
<errorCode>12510</errorCode>
<moduleName>vShield Edge</moduleName>
</error>
Change-Id: If1049035797c525344c46e1ae7f664a7b9daa4a1
- When doing a deletion is a async task we create a new context
to deal with that deletion
- We add some additional logs
Change-Id: I9a707d1292858428ab51295167360bfa26513430
VDR-connected networks were using designated DHCP Edge appliances
to provide metadata.
That was necessary before we introduced option 121 - which can be
used for route injection, which directs metadata traffic towards the
DHCP Edge.
This change removes some redundant code which is supporting metadata
in the old manner.
The patch deprecates supports of older versions of NSX which do not
support insertion of host routes which is required for the change.
Therefore dhcp_force_metadata config parameter has been deprecated.
Change-Id: I6b5e2acf09ce61c87d8ae97471955599cddf320b
Commit If07937636f2312e86ae71d7345053ef72b28e2aa added the logic to
delete edges which are waiting to be created for too long.
This commit does the same also for edges waiting to be updated or
deleted.
Change-Id: I9ca1c525593c28fb9ee7610d2f1dd3cee6f81939
There may be an edge where a back edge is created prior to a edge
being allocated from the pool.
Both of the callbnacks for the edge complete would create the edge
host groups.
This will make sure that we validate that it does not exist prior.
Change-Id: If8c5afed1ce3076e4458b50eb11ffb87e7534e10
Ensure that we treat the locks in the same order with the
delete and create operations.
The fix above result in a deadlcok due to a change in neutron
where ipam delete call port_update. This conflicted with
commit d89eba1a85.
To ensure locking for this we make use of a new lock for the
DHCP interfaces.
Change-Id: I6c3f25ab40247853024560c00d3faa106e5d90b8
When using the exclusive DHCP flag and we have more than one subnet
then make use of the existing edge.
Change-Id: I7b1c779db185b439bf5caf4f6578e7487a78aeff
The DHCP option tftp-server-name is referred to upstream as tftp-server.
This patch supports both names as the same backend option.
Change-Id: Ied936532eafadcd11cb83c51d456509f9d5eedd6
neutron-lib 1.6.0 is out and among other things contains the rehomed
extra_dhcp_opt API definition. This patch switches usage of such
references from neutron to neutron-lib.
Change-Id: I01736dd067b962689a2fe81903a12099dd638e82
Many of the constants from neutron.plugins.common.constants are now in
neutron-lib. This patch switches over to those in neutron-lib.
Change-Id: Ic266440aae034783e5371842ab293da70deeae04
If edge_ha=True and ha_placement_random=True
The user can configure more than 2 edge_host_groups globally or
per availability zone.
In this case 2 of those will be randomly selected for each deployed edge.
Change-Id: Iaa673e5acf78ecdf8cbc942499cce70a2fe0546c