Rebooting neutron node causes all floating IPs on a different neutron node to become unreachable
Issue
- One network node was rebooted. When this node came back online, all of the floating IPs being hosted on one of the other neutron nodes became unreachable.
- Further investigation showed that ovs-vswitchd crashed on the afftected node when the first one came back online. This caused the GRE tunnels to disappear and therefore a network outage for the instances.
Environment
- Red Hat OpenStack Platform 6
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.