The neutron.plugins.ml2.drivers.agent.config module registers options
commonly used by the ml2 agents but in fact it is used only by linux
bridge agent and macvtap agent.
This change makes all ml2 agents use that base module consistently in
individual config modules.
Change-Id: Ib3ec8a8eaf347721bb06f092a0887e62f3a6bffd
I forgot to write a release note when pushing change 834162 [1].
It may be an important change for operators so it's good to have a
release note about that.
[1] https://review.opendev.org/c/openstack/neutron/+/834162
Related-Bug: #1952907
Change-Id: Ie707f461af11357d6eaa004bc98c7eb09a62202f
libvirt-8.4.0-1.el9 caused a regression which resulted into
AttachInterfacesTestJSON and test_hotplug_nic tempest test to
fail, until fixed libvirt version is available in CentOS 9-Stream
repos, let's exclude these tests from jobs running
on CentOS 9-Stream.
Related-Bug: #1979047
Related-Bug: rhbz#2092856
Change-Id: Ia97fc084dcbdfbced0c2cc79f127a97b4d97d1e3
Setting the log-tag line in the haproxy configuration file with the
network or router ID it's possible to filter the logs and make
troubleshooting easier.
This change makes the ovn-metadata-proxy haproxy align with how
ns-metadata-proxy configures it (see 703ff85b82).
Change-Id: I2b3052ecc4f75d127901021c159c5f20eb14a8f8
FIPS jobs reboot nodes after enabling FIPS,
but unbound service takes some time to
be ready and until then DNS resolution fails.
[1] added support to check DNS is ready after
reboot, in order to utilize it we need to
set "nslookup_target" job var.
[1] https://review.opendev.org/c/zuul/zuul-jobs/+/844704
Closes-Bug: #1978938
Change-Id: Ia39753f7cc533ba9959136b95bc3401d2f34cad4
The minimum bandwidth OVS QoS policy is set in the physical bridge
interface, not the port that receives the Neutron QoS policy.
If no other port has a minimum bandwidth rule, the OVS QoS policy
can be unset from this port and deleted. This patch fixes the
port that must be updated.
Closes-Bug: #1977819
Change-Id: I0ebcc85f5ffa9dc29ffedee0c7df1828c85e1576
Fixed "ip route" commands to work with pyroute2 >= 0.6.10, that
introduces APIv2 for link, neighbour and route [1].
[1]bc0f5e2209
Closes-Bug: #1979031
Change-Id: Id2239b6827485a4d466b0916947428ceabef9139
pyroute2 0.6.11 did some further refactoring of their module imports,
so using pyroute2.netlink.some_func() doesn't work anymore. Since we
do "from pyroute2 import netlink" anyway, use the imported module
everywhere.
Also, mock.patch()ing pyroute2.netlink.rtnl.tcmsg.common.tick_in_usec
doesn't work anymore, but also doesn't seem to be necessary for the
tests to succeed, so just drop it.
Closes-Bug: 1978564
Change-Id: Ie06b987f7efbff8057293893e42bee7b682b9ba1
In some cases, the creation of router was called by `l3 plugin`
directly, didn't call API. In this way, the router's request_body
of the `ndp_proxy` service plugin received has no `enable_ndp_proxy`,
this will result in `ndp_proxy` service plugin raise error.
Closes-Bug: #1978519
Related-Bug: #1877301
Change-Id: I413c33970002532d11de5bc2afb145704a8488a9
There was a code that worked around a bug in openvswitch. The bug was
fixed in openvswitch 2.13 and we no longer use that version. It's safe
to be removed now.
Change-Id: I8d0e1dac4f6a6d201e29863cc76db3f1ff8595ae
Signed-off-by: Jakub Libosvar <libosvar@redhat.com>
Port create/update are most time-consuming operations
on subnet creation. As example, in those cases where
several subnets are created over the same network the
response time for those port operations is linearly
increased as the total subnets increases.
This patch improves the number of queries required on port
operations in order to reduce the response time.
Closes-Bug: #1977831
Change-Id: I0fccf36a2035e8f6c2fa8dab0307358da600c8f7
In ``TestPortsV2`` tests, set the
``neutron.plugins.ml2.plugin.MAX_BIND_TRIES`` to 1 to minimize the
number of retries during a failed port binding. That shortens all
test cases execution time if a port binding is executed.
Trivial-Fix
Change-Id: If0473031797984aab5b36c479fcb774e57ff5624
agent's heartbeat_timestamp returns the current time, not
conducive to agent status troubleshooting.
this patch use agent's updated_at as heartbeat_timestamp.
Closes-bug: #1977629
Change-Id: Idf522a20f9735829ee568020bfed46345a95e294
we cannot find this parameter passed in anywhere except for some unit
tests. currently we have used nb_cfg_timestamp[1] as agent updated
time. there are no other scenarios for this parameter. this patch
remove it and update some unit test.
[1] https://review.opendev.org/c/openstack/neutron/+/802834
Closes-bug: #1978035
Change-Id: Ic964b7ddc70988bb1a822b07be6a1be4d197287e