Ask Your Question
2

neutron namespaces not created in all cluster nodes [closed]

asked 2015-04-16 07:39:10 -0500

Fanis gravatar image

updated 2015-04-16 07:39:32 -0500

Dear all, i am facing a serious issue with my openstack setup. I am running openstack Icehouse on Ubuntu 12.0.4.1 and used Mirantis Fuel 5.1.1 to deploy. I have a HA architecture with three controllers/network nodes. Moreover i am using one L3 agent to run two external networks. The above setup was working well for some months now when suddenly some of my users complained that they cannot access their VM's from the internet (via floating IP).

Debugging the problem i found out that not all of my namespaces are created on all of my network/controller nodes. For example i have node-1, node-2 and node-3. When the L3 agent runs on either node-2 or node-3 only one qrouter namespace is created/loaded, when the L3 agent runs on node-1 i have 19 q router namespaces. Same goes for the qdhcp namespaces. When they run on node-1 and node-3 i have 25 qdhcp namespaces. Node-2 loads only 1.

I checked the mysql tables and all networks and routers are on all nodes. Has anybody else faced something like this?? I am in desperate need for help:)

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by omar-munoz
close date 2015-04-21 10:30:23.254453

2 answers

Sort by ยป oldest newest most voted
1

answered 2015-04-21 07:17:54 -0500

Fanis gravatar image

Ok i have solved my issue. The thing was that although qdhcp namespaces exist and are "copied" via the cluster the router namespaces do not. Due to the problem i had faced (power issues with controllers) some of the qrouter namespaces did not "switch" L3 agent and they remained on ones that did not run. I transfered the qrouter namespaces on the running L3 agent and they work fine now.

Thanks for taking the time to answer my question Omar.

edit flag offensive delete link more
1

answered 2015-04-16 10:41:58 -0500

Are you using crm to move the resources, can you run 'crm status' and post the output. Have you looked at your /var/log/neutron-all.log for more information.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2015-04-16 07:39:10 -0500

Seen: 666 times

Last updated: Apr 21 '15