Is this normal behavior?

2 node cluster, Node A and Node B, running 2 resource groups- 1 for each
node and mutually exclusive of each other.
Resource group A has alot of IPs and runs ipvsadm and ldirectord. 
Resource group B has nothing - yet. This group will contain all that is
needed for this server to be a realserver(lvs term).

I have 4 additional hosts that are dedicated realservers- no heartbeat.

Node A is running Resource Group A
Node B is running Resource Group B

While do some initial setup of my cluster today, I had some problems
reaching a load balanced ip address. While solving THAT problem(I solved
it btw), I did some digging and I found the following:

Node B has almost all the IP addresses that Node A has. 

There are 2 differences- 
1-    the IPs on host B are bound to the *lo* interface and are
contained in Resource Group A.
2-    some IPs on host A in Resource Group A are NOT on Host B at all

Is this normal? 

Are ALL the IPs on Host A supposed to be on Host B, just bound to the LO
interface?

Does this mean that I could add my 4 additional servers to this cluster
and that they could all be potential lvs directors,(one at a time, of
course), while the rest of them have the load balanced IPs bound to the
LO interface? I do this manually on my other realservers now.

I hope I am making sense.

Thanks,
Dave


_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to