Comment 3 for bug 1593354

So I'm unable to reproduce this on master (didn't try on mitaka). Basically I set up 2 dvr_snat nodes and performed the reboot as specified in comment #2 on the master node - the failover of the other node was immediate.

Also note that starting from Newton, all nodes have the snat- namespace (and interfaces, only without IPs) to allow for faster failover).

Hao, can you please provide the keepalived version you used in your deployment?