DVR+HA routers all answering to ping on private interface

When using HA routers, the qr-xzy link interface in qrouter-namespace is set UP/DOWN based on keepalived status change.

When using DVR+HA routers, with 2 network nodes (in dvr_snat mode), the qr-xzy interface in qrouter-namespace is NOT managed anymore by keepalived. In fact, keepalived is running in a snat-namespace and have no access to this qr-xzy interface.

The result is that the qr-xyz link interface is always UP in qrouter-namespace, even if the router in in standby/backup mode.

The result is that, if any other equipment (e.g. ironic node) in the private network is trying to ping the qr-xyz IP address (e.g. 192.168.43.1), then both routers are answering:
$ arping -c1 192.168.43.1
ARPING 192.168.43.1
60 bytes from fa:16:3f:67:97:6a (192.168.43.1): index=0 time=634.700 usec
60 bytes from fa:16:3f:dc:67:91 (192.168.43.1): index=1 time=750.298 usec

--- 192.168.43.1 statistics ---
1 packets transmitted, 2 packets received, 0% unanswered (1 extra)

Note 1: a topic was starting on openstack-discuss regarding this issue:
https://lists.openstack.org/pipermail/openstack-discuss/2022-December/031480.html

Note 2: this bug describes only the "snat" (network node) part of the issue. DVR is also running on hypervisors, this will eventually be discussed in another bug.