neutron restart the service result in part of the equipment tag becomes a problem 4095

restart control node (network node) the vxlan mode neutronservice will cause ovs ha qr qg routers and other equipment for the tag device 4095 cause the tag can not communicate on the

 

Now Problem Solution

 
  • Modify database information

 

4095 port neutron.ml2_port_bindingsstatus table must bebind_failed

 
  1. update neutron.ml2_port_bindings set vif_type='ubound' where port_id like {PORTID}
 
  • Restart openvswitch neutron-openvswitch-agentsolve

 

 

 

The problem is that the community openstack 4095 bug, the community has been discussed in detail, the fundamental reason is that neutron-server service than neutron-openvswitch-agent due to start soon
1. The server is down, the service on this server is marked as neutron down state
2. After the server is powered on, neutron server service first started, check the agent found the service did not start, the ports are labeled bind failed state.
3. agent after the start, we found that these ports are down state. . . Then put them into a 4095 marker. . .
 
 
P version of this issue has been resolved, but the code changes more, we do not deserve to spend too much energy to backport come back, we still have to test to solve it by managing the service startup sequence.

 

Reproduced in: https: //www.cnblogs.com/lsw-blogs/p/11077650.html

Guess you like

Origin blog.csdn.net/weixin_33775572/article/details/94078666