6

I have two nodes running keepalived, and two VIP, e.g.

Node 1      Node 2
VIP1        VIP2

So in each node, I have two definition of vrrp_instance, so I assume the two vrrp_instance in my keepalived.conf suppose to be unique, but they can be the same values on another node, right?

Ryan
  • 5,341
  • 21
  • 71
  • 87
  • If they're the same instance, they should have the same instance ID. If they're different instances, they should have different instance IDs. An instance that only existed on one node wouldn't be very useful. – David Schwartz Jan 28 '13 at 14:14

1 Answers1

11

Different VRRP instances should have different virtual_router_id values. Same VRRP instances should have the same value.

You can read the following from man keepalived.conf:

# arbitary unique number 0..255
# used to differentiate multiple instances of vrrpd
# running on the same NIC (and hence same socket).
virtual_router_id 51

To summarize, you need to have the same value on the members of same cluster. If you have another cluster, its members should have another value. The virtual_router_id should be unique per VRRP cluster.

HopelessN00b
  • 53,385
  • 32
  • 133
  • 208
Khaled
  • 35,688
  • 8
  • 69
  • 98
  • 1
    Hi, in my network I have any set of keepalived nodes which has the dupltcated virtual_router_id, but I have tested the function is ok, execept there are warning log syslog, does it matter? Also, is it recommended to use the same string for both vrrp_instance and virtual_router_id? Thanks – Ryan Jan 30 '13 at 04:58