Open Stack

Very annoying situation, every time it really hard to debug.

My debug experience so far:

Easy:
1. Check every /var/log/upstart/neutron*.log file. Sometime stupid typo 
in configs cause problems.
2. Check access group for instance.

In depth:
Check ovs-ofctl dump-flows br-tun on compute where instance is running. 
Proper configuration should contain lines like this:

cookie=0x0, duration=216.991s, table=21, n_packets=19, n_bytes=2326, 
idle_age=144, priority=1,dl_vlan=3 
actions=strip_vlan,set_tunnel:0x1,output:6,output:5,output:2,output:3,output:4
(table=21, and set_tunnel, output - is important).

I create a very crude utility to probing gre tunnels: 
https://github.com/amarao/gre-tun-probe to check if everything is fine.

One more hint: Ubuntu cut off neutron-ovs-cleanup script. Run it and 
reboot compute node, may be it will helps.


On 29.01.2014 17:25, Alvise Dorigo wrote:
> Hello,
> I've a distributed installation of OpenStack Havana on 3 different nodes:
>
>  1. Controller node: nova-api + nova-cert + nova-scheduler + 
> nova-consoleauth + nova-novncproxy + nova-conductor + keystone + 
> glance + neutron server + Dashboard
>
>  2. compute node: nova-compute
>
>  3. network node: l3-agent + dhcp-agent + metadata-agent
>
> Of course, on all of them, the l2-agent (neutron-openvswitch-agent) is 
> also running.
>
> From the dashboard I created a simple cirros image; then I launched it 
> and attached to an non-external network (created in the Admin project, 
> and having its own subnet and dhcp enabled). Nova assigned an IP to 
> the VM, but actually when the VM's operating system bootstrapped it 
> couldn't obtain from the DHCP the assigned IP.
>
> I monitored the br-int interface on the compute node and on the 
> network node, using "tcpdump -i br-int".
>
> I saw the packet traffic going through the two br-int interfaces (from 
> compute node to network node). But when I monitored the tapXYZ... in 
> the dhcp network namespace (in the network node) I didn't see any 
> packet traversing it:
>
>     ip netns exec qdhcpXYZ... "tcpdump -i tap37fgw74frwgc"
>
> ... nothing ... :-(
>
> So, it seems that the DHCP client's request correctly originates from 
> the VM, and is correctly trasmitted to the network node by mean of the 
> data network the two br-int interfaces are connected to; but for some 
> reason the dhcp-agent doesn't receive anything, as proved by the null 
> traffic on its tap.
>
> Any idea about what and how I could more deeply investigate ?
>
> thanks,
> Alvise
>
>
> _______________________________________________
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20140129/7022070f/attachment.html>