VIP does not fail over when public interface is down
bentwDec 31 2009 — edited Jan 6 2010Dear all,
This is my enviroment settings:
OS: RHEL4.8 IA64
Clusterware : 10.2.0.4
RAC Database : 10.2.0.4
Node1: Bond0--including eth0 & eth1--> for Public IP using IP 193.X.X.X
Bond1--including eth2 & eth3--> for RAC interconnect using IP 172.16.X.X
Node2: Bond0--including eth0 & eth1--> for Public IP using IP 193.X.X.X
Bond1--including eth2 & eth3--> for RAC interconnect using IP 172.16.X.X
When I pulled out the lines in both eth0ð1 on Node1, I use the command " cat /proc/net/bonding/bond0 " to ensure that bond0 is offline, but I CANT see my VIPinNode1 failover to Node2; and when I type "$CRS_HOME/bin/crs_stat -t" on both Node1 and Node2, it shows that the VIPinNode1 is "STILL IN NODE1"....... But the right process should the VIP be failed over from Node1 to Node2, isn't it ?
I can't figure out what's wrong and why the VIP do not fail over. Is there any log I can check for more clues?
By the way, when I totally shutdown Node1, the vip will correctly failover to Node2.
Can any one help me to find out why VIP does not failover when public network interface is down ?
Any help would be appreciated.