Not able to ping to new VM if I don't use vnc. (v3.8)

Simply put, when I put GRAPHICS=[TYPE=“vnc” ] in the VM template, I am able to ssh/ping to the VM. If I don’t use VNC, I am not able to ping/ssh to the vm.
Not sure where the problem is.

FYR:
OS: CentOS6.4 (on host m/c and VM)
Opennebula Version: 3.8

Onevnet show:

[root@server]# onevnet show 2
VIRTUAL NETWORK 2 INFORMATION
ID : 2
NAME : New_Net2
USER : oneadmin
GROUP : oneadmin
CLUSTER : -
TYPE : RANGED
BRIDGE : br0
VLAN : Yes
PHYSICAL DEVICE: eth0
VLAN ID :
USED LEASES : 2

PERMISSIONS
OWNER : um-
GROUP : —
OTHER : —

VIRTUAL NETWORK TEMPLATE
NETWORK_MASK=“255.255.192.0”

RANGE
IP_START : 172.31.4.21
IP_END : 172.31.4.31

USED LEASES
LEASE=[ IP=“172.31.4.21”, MAC=“02:00:ac:1f:04:15”, USED=“1”, VID=“144” ]
LEASE=[ IP=“172.31.4.22”, MAC=“02:00:ac:1f:04:16”, USED=“1”, VID=“133” ]

And Onevm show:

[root@nfsserver 145]# onevm show 144
VIRTUAL MACHINE 144 INFORMATION
ID : 144
NAME : check
USER : oneadmin
GROUP : oneadmin
STATE : ACTIVE
LCM_STATE : RUNNING
RESCHED : No
HOST : c2
START TIME : 07/27 10:44:47
END TIME : -
DEPLOY ID : one-144

VIRTUAL MACHINE MONITORING
USED CPU : 0
NET_RX : 0K
USED MEMORY : 0K
NET_TX : 0K

PERMISSIONS
OWNER : um-
GROUP : —
OTHER : —

VIRTUAL MACHINE TEMPLATE
CONTEXT=[
DISK_ID=“1”,
ETH0_MASK=“255.255.192.0”,
GRAPHICS=“vnc”,
TARGET=“hdb” ]
CPU=“1"
DISK=[
CLONE=“YES”,
DATASTORE=“default”,
DATASTORE_ID=“1”,
DEV_PREFIX=“hd”,
DISK_ID=“0”,
DRIVER=“qcow2”,
IMAGE=“002”,
IMAGE_ID=“27”,
READONLY=“NO”,
SAVE=“NO”,
SOURCE=”/var/lib/one/datastores/1/d492d75171e4613fd660f09616ed7496",
TARGET=“hda”,
TM_MAD=“shared”,
TYPE=“FILE” ]
INSTANCE_TYPE="small"
MEMORY="1024"
NAME="check"
NIC=[
BRIDGE=“br0”,
IP=“172.31.4.21”,
MAC=“02:00:ac:1f:04:15”,
NETWORK=“New_Net2”,
NETWORK_ID=“2”,
PHYDEV=“eth0”,
VLAN=“YES” ]
VCPU="1"
VMID=“144”

VIRTUAL MACHINE HISTORY
SEQ HOST REASON START TIME PROLOG_TIME
0 c2 none 07/27 10:44:55 0d 00h12m06s 0d 00h08m16s

My one directory is shared between host machine and nebula server.