Description of problem: in a RHOSP 16.2.4 - NonDVR l3-ha OVS environment floating ips stopped working (not reachable) suddenly. Workaround was about deleting and re-creating "affected" routers. Currently, number of no-working routers keeps increasing. Version-Release number of selected component (if applicable): RHOSP 16.2.4 - NonDVR l3-ha OVS environment Additional info: more logs and info will be attached as comments asap.
Hi Team, As the customer has closed the case now and we are unable to reproduce, I will be closing this bugzilla now. If you suspect you are hitting this same issue please check router status on all 3 controllers with something like [1] If you see the route status for a given router is master on all 3 controllers, you can test recreating it and checking the status once more. Some things that may be relevant: - neutron l3-agent-list-hosting-router <UUID> output were void for the failing routers - Environment was recently updated from 13.0.16 to 16.2.4 - Issue didn't reoccur after routers were recreated. If your issue matches some of the above, please reopen this bugzilla. Thanks! [1] for i in $(ls *2023-06-06*/*/var/lib/neutron/ha_confs/ | grep -v '\.' |sort -u); do echo -n "$i "; rg master *2023-06-06*/*/var/lib/neutron/ha_confs/$i/state | wc -l | tr -d '\n' ;echo -n " "; rg backup *2023-06-06*/*/var/lib/neutron/ha_confs/$i/state | wc -l ;done