Add configuration option for Neutron OVN Metadata workers
The Neutron default of number of CPUs / 2 is way too high for the
Neutron OVN Metadata service and puts an unneccessary toll on the
OVN Southbound database.
Add a new configuration option to control this which defaults to
2.
Note that we do not use the default worker-multiplier handling
code that many other of our charms use because the main use case
for that code is to handle the use case where API services run in
containers.
The Neutron OVN Metadata service runs directly on the host.
Merge after openstack-charmers/charm-layer-ovn#21
Change-Id: Ic335a552dc9afcaab026c534499e4070846851eb
Closes-Bug: #1893656
(cherry picked from commit e06b578b4580c0f46fe38f8937934a6ff72b97a3)
At present linux bonds will be ignored when mentioned directly
in configuration options parsed by the ``BridgePortInterface``
data structure.
If a mac address of a physical interface making up the bond is
listed the code correctly finds the bond name and adds that to
the bridge.
Virtual interfaces are ignored when then map is built, linux bonds
are virtual interfaces.
This patch fixes the problem by checking physical interfaces for
bond membership while building the map. If an interface is member
of a bond add the bond name to the map.
Previously the bond membership check was only done when looking
up interface name from mac after building the map.
- Adds groovy to the series in the metadata
- Classic charms: sync charm-helpers.
- Classic ceph based charms: also sync charms.ceph
- Reactive charms: trigger a rebuild