Problem Description: Clearly explain the issue.

When ipam is configured (e.g. ovn-nbctl lsp-set-addresses swp1 "02:00:00:00:00:01 dynamic"), ovn-northd assigns an IP address.

However, northd might assign an already used IP address and write it to NB. In the following iteration (on notification of NB change), northd assigns a proper address.

So a wrong IP might be used by ovn-controller, for the duration of a northd loop.

 Impact Assessment: Describe the severity and impact (e.g., network down,availability of a workaround, etc.).

 Software Versions: Specify the exact versions in use (e.g.,openvswitch3.1-3.1.0-147.el8fdp).

ovn/main

  Issue Type: Indicate whether this is a new issue or a regression (if a regression, state the last known working version).

This issue was introduced by 'northd: Incremental processing of LSPs IPAM in "northd" node.'

 Reproducibility: Confirm if the issue can be reproduced consistently. If not, describe how often it occurs.

In common setups, difficult to reproduce - as only visible between two iterations of northd. But visible in NB logs. Also reproduced through the following unit test (a modified ipam router ports unit test).

 Reproduction Steps: Provide detailed steps or scripts to replicate the issue.

OVN_FOR_EACH_NORTHD([
AT_SETUP([ipam router ports])
ovn_start
check ovn-nbctl ls-add sw
check ovn-nbctl set logical_switch sw other-config:subnet=192.168.1.0/24
for i in 2 3 4; do
    check ovn-nbctl lr-add ro$i
    check ovn-nbctl lsp-add sw swp$i
    sleep_northd
    check ovn-nbctl lsp-set-addresses swp$i "02:00:00:00:00:0$i dynamic"
    sleep_nb
    wake_up_northd
    sleep 1
    sleep_northd
    wake_up_nb
    sleep 1
    cidr=$(ovn-nbctl get logical_switch_port swp$i dynamic_addresses |cut -f2 -d' '|cut -f1 -d\")
    wake_up_northd
    check ovn-nbctl lrp-add ro$i rop$i 02:00:00:00:00:0$i $cidr/24 -- set logical_switch_port swp$i type=router options:router-port=rop$i addresses=router;
    AT_CHECK_UNQUOTED([ovn-nbctl get logical_router_port rop$i networks], [0], [[["192.168.1.$i/24"]]
])
done
ovn-nbctl list logical_switch_port
ovn-nbctl list logical_router_port
AT_CLEANUP
])

with the following addition in ovn-macros.at

sleep_nb() {
  echo NB going to sleep
  AT_CHECK([kill -STOP $(cat ovn-nb/ovsdb-server.pid)])
  on_exit "test -e ovn-nb/ovsdb-server.pid && kill -CONT $(cat ovn-nb/ovsdb-server.pid)"
}
wake_up_nb() {
  echo NB waking up
  AT_CHECK([kill -CONT $(cat ovn-nb/ovsdb-server.pid)])
} 

Expected Behavior: Describe what should happen under normal circumstances.

Wrong (already used) IP should not be re-used.

 Observed Behavior: Explain what actually happens.

 Troubleshooting Actions: Outline the steps taken to diagnose or resolve the issue so far.

 Logs: If you collected logs please provide them (e.g. sos report, /var/log/openvswitch/* , testpmd console)