Ask Your Question
0

quantum single flat network, but cannot ping instances

asked 2013-03-18 06:03:09 -0600

digitalwonk gravatar image

Hello,

I setup a single flat network according to the following documentation demo:

http://docs.openstack.org/trunk/openstack-network/admin/content/app_demo_flat.html (http://docs.openstack.org/trunk/opens...)

However, instances cannot be pinged from either the compute nodes or the quantum network node (the node with the dhcp agent). Here is a paste from 'quantum net-show', 'quantum subnet-show', 'nova list', 'ip link', and 'ovs-vsctl show'.

http://paste.openstack.org/show/33935/

I created three instances, and they all seem to be assigned the right ip address, within the 1.1.1.0/24 network. The nova compute and quantum network nodes are also on the 1.1.1.0/24 network (they have ips assigned to br-eth0 after bridge setup). Can anyone provide tips as to what to check?

Thank you, Ed

edit retag flag offensive close merge delete

6 answers

Sort by ยป oldest newest most voted
0

answered 2013-03-18 06:39:10 -0600

Can you check via VNC if VMs get IP addresses on their interfaces?

edit flag offensive delete link more
0

answered 2013-03-18 14:22:28 -0600

digitalwonk gravatar image

No, it does not from the console.logs (an excerpt):

[ OK ]^M Bringing up loopback interface: [ OK ]^M Bringing up interface eth0:
Determining IP information for eth0... failed. [FAILED]^M Starting system logger: [ OK ]^M Starting kernel logger: [ OK ]^M iscsid (pid 1231) is running... Setting up iSCSI targets: iscsiadm: No records found [ OK ]^M

edit flag offensive delete link more
0

answered 2013-03-18 14:29:31 -0600

digitalwonk gravatar image

I just checked the ovs-ofctl on the quantum network node:

ovs-ofctl show br-int

OFPT_FEATURES_REPLY (xid=0x1): ver:0x1, dpid:0000d6b390e5324c n_tables:255, n_buffers:256 features: capabilities:0xc7, actions:0xfff 1(int-br-eth0): addr:ee:ae:e4:44:7d:3b config: 0 state: 0 current: 10GB-FD COPPER 3(tapdd91d332-c6): addr:e7:00:00:00:00:00 config: PORT_DOWN state: LINK_DOWN LOCAL(br-int): addr:d6:b3:90:e5:32:4c config: 0 state: 0 OFPT_GET_CONFIG_REPLY (xid=0x3): frags=normal miss_send_len=0

And it seems to think that the tap is down. Is that ok? If not, how can I get the link state up? The dhcp agent is running or at least there are two processes for it:

nobody 24743 0.0 0.0 28820 1096 ? S Mar17 0:00 dnsmasq --no-hosts --no-resolv --strict-order --bind-interfaces --interface=tapdd91d332-c6 --except-interface=lo --domain=openstacklocal --pid-file=/var/lib/quantum/dhcp/6f12748f-2cae-4746-815d-66338e267b2f/pid --dhcp-hostsfile=/var/lib/quantum/dhcp/6f12748f-2cae-4746-815d-66338e267b2f/host --dhcp-optsfile=/var/lib/quantum/dhcp/6f12748f-2cae-4746-815d-66338e267b2f/opts --dhcp-script=/usr/bin/quantum-dhcp-agent-dnsmasq-lease-update --leasefile-ro --dhcp-range=set:tag0,1.1.1.0,static,120s root 24744 0.0 0.0 28792 244 ? S Mar17 0:00 dnsmasq --no-hosts --no-resolv --strict-order --bind-interfaces --interface=tapdd91d332-c6 --except-interface=lo --domain=openstacklocal --pid-file=/var/lib/quantum/dhcp/6f12748f-2cae-4746-815d-66338e267b2f/pid --dhcp-hostsfile=/var/lib/quantum/dhcp/6f12748f-2cae-4746-815d-66338e267b2f/host --dhcp-optsfile=/var/lib/quantum/dhcp/6f12748f-2cae-4746-815d-66338e267b2f/opts --dhcp-script=/usr/bin/quantum-dhcp-agent-dnsmasq-lease-update --leasefile-ro --dhcp-range=set:tag0,1.1.1.0,static,120s

edit flag offensive delete link more
0

answered 2013-03-18 14:41:42 -0600

digitalwonk gravatar image

Ok, so, apparently the tap device cannot be found, according to the dhcp-agent.log on the network node:

Stderr: 'Cannot find device "tapdd91d332-c6"\n' 2013-03-18 07:37:52 ERROR [quantum.agent.dhcp_agent] Unable to enable dhcp. Traceback (most recent call last):

How do can I make it be found?

edit flag offensive delete link more
0

answered 2013-03-19 17:53:48 -0600

digitalwonk gravatar image

I actually rebooted the quantum network node (which has the dhcp agent) and now the tap interface is present, but strangely I cannot access or ping the quantum server (on the nova controller) from the network node. The quantum server also has the rabbitmq service and quantum database. I will provide more information after I debug a bit more.

edit flag offensive delete link more
0

answered 2013-03-19 19:55:30 -0600

danwent gravatar image

it seems likely that the quantum-ovs-cleanup script is not being run on reboot.

you could try running it manually, and then restarting the dhcp agent.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2013-03-18 06:03:09 -0600

Seen: 54 times

Last updated: Mar 19 '13