Ask Your Question
0

Liberty RDO "Neutron with existing external network"

asked 2016-01-14 05:25:43 -0500

JariK gravatar image

updated 2016-01-25 05:08:56 -0500

I've been trying to do "Neutron with existing external network" install by using the latest centos 7 + RDO. When booting a Fedora 23 instance, I'm stuck with this (cirros has similar symptoms as well):

ci-info: +--------+------+-----------+-----------+-------+-------------------+
ci-info: | Device |  Up  |  Address  |    Mask   | Scope |     Hw-Address    |
ci-info: +--------+------+-----------+-----------+-------+-------------------+
ci-info: |  lo:   | True | 127.0.0.1 | 255.0.0.0 |   .   |         .         |
ci-info: |  lo:   | True |     .     |     .     |   d   |         .         |
ci-info: | eth0:  | True |     .     |     .     |   .   | fa:16:3e:34:71:a3 |
ci-info: | eth0:  | True |     .     |     .     |   d   | fa:16:3e:34:71:a3 |
ci-info: +--------+------+-----------+-----------+-------+-------------------+
[WARNING]: Calling 'http://169.254.169.254/2009-04-04/meta-data/instance-id' failed [0/120s]: request error [('Connection aborted.', OSError(101, 'Network is unreachable'))]

What I did was:

  1. Follow instructions in https://www.rdoproject.org/networking/neutron-with-existing-external-network/ (https://www.rdoproject.org/networking...) (except added "--os-horizon-ssl=y --ntp-servers=..." to packstack and modified /etc/neutron/plugins/ml2/openvswitch_agent.ini and /etc/neutron/plugins/ml2/ml2_conf.ini as instructed in https://ask.openstack.org/en/question/86601/neutron-with-existing-external-network/ (https://ask.openstack.org/en/question...) (except replaced physnet1 by extnet)).

  2. Fixed the quota problem: https://ask.openstack.org/en/question/87002/cinder-volume-quotas-broken-for-rdo-install-in-centos/ (https://ask.openstack.org/en/question...)

  3. Tried to create a Fedora/Cirros instance: https://www.rdoproject.org/networking/running-an-instance-with-neutron/ (https://www.rdoproject.org/networking...)

I can ping the first address of the floating IP range, but not the assigned one.

# ifconfig
br-ex: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.122.50  netmask 255.255.255.0  broadcast 192.168.122.255
        inet6 2001:db8::7a2b:cbff:fe47:f037  prefixlen 64  scopeid 0x0<global>
        inet6 fe80::7a2b:cbff:fe47:f037  prefixlen 64  scopeid 0x20<link>
        ether 78:2b:cb:47:f0:37  txqueuelen 0  (Ethernet)
        RX packets 17433  bytes 2293095 (2.1 MiB)
        RX errors 0  dropped 15  overruns 0  frame 0
        TX packets 2751  bytes 2667610 (2.5 MiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

em1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::7a2b:cbff:fe47:f037  prefixlen 64  scopeid 0x20<link>
        ether 78:2b:cb:47:f0:37  txqueuelen 1000  (Ethernet)
        RX packets 18641  bytes 2452312 (2.3 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 3904  bytes 2760240 (2.6 MiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 0  (Local Loopback)
        RX packets 286112  bytes 44247917 (42.1 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 286112  bytes 44247917 (42.1 MiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

qbrd0bd198a-f7: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::643b:4ff:fe60:bbfe  prefixlen 64  scopeid 0x20<link>
        ether 6a:cc:4a:ac:f1:f7  txqueuelen 0  (Ethernet)
        RX packets 28  bytes 5260 (5.1 KiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 8  bytes 648 (648.0 B)
        TX errors 0 ...
(more)
edit retag flag offensive close merge delete

Comments

I installed kilo on a similar machine with success. I compared ifconfig etc. outputs between kilo and liberty without seeing any significant differences. However, in liberty I can't access default router inside qdhcp namespace. Logs added above. Any idea what's wrong?

JariK gravatar imageJariK ( 2016-01-21 02:36:55 -0500 )edit

PS. Are rdo debugging instructions really up-to-date (https://www.rdoproject.org/troubleshooting/networking/) (https://www.rdoproject.org/troublesho...) At least in my case br-int is down by default so I can't tcpdump it.

JariK gravatar imageJariK ( 2016-01-21 02:40:32 -0500 )edit

2 answers

Sort by ยป oldest newest most voted
0

answered 2016-01-21 03:56:36 -0500

Prateek K gravatar image

Based on what I read in the post I see that the interface on the VM is not getting the IP from dhcp server running for your internal network..

Do s ps -aef |grep dnsmasq on the neutron node and check weather a dhcp server is running for the internal network or not.

If the dhcp server is running check the flows in both compute and neutorn node of OVS via command ovs-ofctl dump-flows <bridge> ....

If the flows are also fine then do a tcpdump on the interfaces and see weather neutorn node is actually receving the DHCP broadcast message from teh VM or not.

edit flag offensive delete link more

Comments

Thanks for the answer! Have you been able to install liberty using RDO "Neutron with existing external network" yourself? For me, those instructions work for Kilo, but not for Liberty so it would be helpful to know if the instructions are broken or am I doing something wrong?

JariK gravatar imageJariK ( 2016-01-25 05:15:35 -0500 )edit

dnsmasq is up and running. I'm still trying to understand ovs-ofctl output, but I added it above in case anyone can help interpreting it. Seems that dnsmasq is listening on tapc... interface in qdhcp namespace and tcpdump doesn't show any dhcp there so obviously it is somehow disconnected.

JariK gravatar imageJariK ( 2016-01-25 05:28:53 -0500 )edit

Too much to see :P ... Can we setup a call where we can debug this further.... It will be great if we can have a team-viewer session... My email id is prateek.khushalani@gmail.com ...

Prateek K gravatar imagePrateek K ( 2016-01-28 01:04:44 -0500 )edit
0

answered 2016-02-25 09:03:56 -0500

JariK gravatar image

Thanks for your ideas and support! I decided to start from scratch one more time and now everything works. Either the software or the instructions have been updated because the process was smooth (previously, one of the files mentioned in instructions didn't exist). May be I avoided some mistakes too...

Anyway, comparing current situation and previous, at least tags on br-int are different in "ovs-vsctl show" output. Previously, they were 4095 and 2, now all are 1.

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: 2016-01-14 05:25:43 -0500

Seen: 406 times

Last updated: Feb 25 '16