Created attachment 598431 [details]
engine log

Description of problem:
rhevm 3.1 with cluster 3.0: host doesn't move to non-operational due to missing required network.
This can cause some critical issues with VM connectivity and migration

Steps to Reproduce:
1. create required network "sw2"
2. attach to host eth2
3. move eth2 down (unplug or ifdown: as long as operstat=down)
  
Actual results:
host will stay up

Expected results:
host should move to non-operational after 60 sec

* vdsm does show that eth2 status is down:
# vdsClient 0 getVdsStats
'eth2': {'macAddr': '', 'name': 'eth2', 'txDropped': '0', 'rxErrors': '0', 'txRate': '0.0', 'rxRate': '0.0', 'txErrors': '0', 'state': 'down', ...

# brctl show
bridge name     bridge id               STP enabled     interfaces
rhevm           8000.001517a76a4c       no              eth0
sw1             8000.001d0968714e       no              eth2

[root@silver-vdsa ~]# virsh -r net-list
Name                 State      Autostart
-----------------------------------------
vdsm-rhevm           active     yes       
vdsm-sw1             active     yes