Why does Pacemaker consider removing IP from an interface as a mistake?

Question

High availability

Centos
pacemaker 1.1.10
corosync 1.4.1
Describes the standard IPaddr2 resource. If the interface on which the ip is hanging falls, corosync silently transfers it to the next node, but if you say ip ad 10.0.0.1/24 dev eth0, the address will be transferred, but
Failed actions will be reflected in crm_mon:
VIP_monitor_30000 on zvm1 'not running' (7): call=33, status=complete, last-rc-change='Thu Feb 13 06:50:24 2014', queued=0ms, exec=
0ms the agent code contains such an error code.
Should pacemaker show failed action in this case?

Answer 1

VIP_monitor_30000 on zvm1 'not running' (7):
Indicates that the resource monitoring checked the status and saw that the interface is down (code 7=not running, 0=running, rest=1-6 errors) and transferred the IP to another node. With the right settings, you should not even see this inscription. And it turns out that he has monitored the state and does not know what to do next.
I think it doesn't reroll the virtual ip - it just doesn't get deleted when the node changes. Those. once it was assigned, and then it is already available due to the availability of the IPaddr2 resource.
PS Little data.