ERROR nova.rpc.common [-] Timed out waiting for RPC response: timed out

From the logs it looks like nova-compute sent an RPC request to nova-network, which did not complete. So rpc.timeout on nova-compute logs makes sense.

But why nova-network itself has RPC.timeout error is unclear from nova-network logs. Do you see any other errors in nova-network log ?

Which network manager are you using ? (May not be directly relevant to the error from logs .. but may be useful)

my Controller Node have tow NICS
use FlatDHCPManager for nova-network.

# NETWORK
network_manager=nova.network.manager.FlatDHCPManager
force_dhcp_release=True
dhcpbridge_flagfile=/etc/nova/nova.conf
firewall_driver=nova.virt.libvirt.firewall.IptablesFirewallDriver
my_ip=PUBLIC_IP
public_interface=eth1
vlan_interface=eth0
flat_network_bridge=br100
flat_interface=eth0
fixed_range=10.0.0.0/24
multi_host=True
send_arp_for_ha=True

after recently test I found , where I delete fixed_ip
# nova-manage network delete 10.0.0.0/24
then instanc launch at Node1 successfully.
 Tenant Host Instance Name IP Address Size Status Task Power State Actions
 demo Node1 winxp-Node1 512MB RAM | 1 VCPU | 0 Disk Active None Running
though it hvae get fixed_ip

so when I allcote fixed ip at Control Node which run nova-network

# nova-manage network create private --fixed_range_v4=10.0.0.0/24 --num_networks=1 --bridge=br100 --bridge_interface=eth0 --network_size=256 --multi_host=T

then create a new instance meets following problems:

 Tenant Host Instance Name IP Address Size Status Task Power State Actions
 demo Node1 winxp-Node1 512MB RAM | 1 VCPU | 0 Disk Error Networking No State
 demo cloud winxp-cloud 10.0.0.2 512MB RAM | 1 VCPU | 0 Disk Active None Running

please see details of log

nova-network.log on Controll Node
http://pastebin.com/8HBULmLP

nova-compute.log on Node1
http://pastebin.com/Ksh05duC

I had a similar problem.

In my case (which I believe is exactly the samesituation here) , it was due to the fact that because of some unclear misconfiguration, nova-network did not start the dnsmasq daemon beside it is configured using a FlatDHCP manager.

So the timeout comes from the nova-network daemon waiting for an answer from a dnsmasq that nobody started.

Of course, as usual with OpenStack, there absolutely no meaningful error message anywhere...

About the misconfiguration it *might* be due to the fact I created a multi-host network at first (following one of the examples in the documentation). I'm not sute I have changed anything else. But now, nova-network do start dnsmasq, and at last, I can boot an image.

douardda (david-douard), how did you manage to make dnsmasq work? I have the same problem, I can only start instances without networking (not even in the main node). Did you have to install anything else?

Thank you,