I run a keepalived failover configuration. If a failover happens the backup server gets unresponsive (console & network) all the VIPs gets unresponsive too. When the master comes back the situation gets back to normal after a minute or so - it seems the backup server does not get online.
Keepalived serves 570 VIPs. In a 21 masked network. for every "C"-class range within there is a vrrp instance. Currently 3 vrrp instances with unicast_peer set.
I have tried to use virtual_ipaddress_excluded. But without any change in behavior.
Is there anyone out there that could help or uses keepalived with such many VIPs?
Also the network guys tell me that they detect a broadcast storm in the network when a failover happens. The same configuration runs in other networks with less VIPs and they run fine.
Many Thanks
