[Openstack] [openstack][neutron] security_group_rules_for_devices RPC timeout & vm network fail

sosogh sosogh at mail.com
Wed Aug 31 12:10:39 UTC 2016


we are running Havana (2013.2.3 )(provider network/ovs bridge/dhcp/dnsmasq)。
It is running well in the past .
But today something broken , when creating a new vm , the vm was created successfully , but  vm network failed.

I see "cloud-init-nonet waiting 120 seconds for a network device." via nova console-log  vmid.
vm did not get ip via dhcp.
I tried to ifconfig the vm manually via vnc , but it did not work either.  

There is  error log in  /var/log/neutron/openvswitch-agent.log:
ERROR neutron.plugins.openvswitch.agent.ovs_neutron_agent [-] Error in agent event loop
Timeout while waiting on RPC response - topic: "q-plugin", RPC method: "security_group_rules_for_devices" info: "<unknown>"


There are lots of message which are not comsumed in q-plugin queue in rabbitmq.
The numbers of message  in q-plugin queue is  getting more and more .
root at rise-mq-27:~# rabbitmqctl list_queues  -p /13fa9ce55  name   durable messages |grep q-plugin 
q-plugin        false   135269


My questions are :
1. Does the q-plugin lead to the vm network failture?
2. if so , who are the comsumer of q-plugin , how I can make them comsume again?
3. it is said that it will skip "q-plugin" RPC timeout issue if disabling  security group .
   I've read this post https://ask.openstack.org/en/question/9592/disabling-security-groups-in-horizon/
   and wonder whether  I should edit both(nova and neutron) configuration files only in HV 
   or also in nova/neutron server (controller)?

Thank you .




sosogh
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20160831/5b121f39/attachment.html>


More information about the Openstack mailing list