multiple dhcp ports on a subnet, only one dhcp-agent

Bug #1053597 reported by dan wendlandt
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
neutron
Won't Fix
Undecided
Unassigned

Bug Description

in the setup below, noticed that subnet "c37d9f81-eede-4ac6-9397-d10c240a3d11" some how got two DHCP address assignments. There is only a single host running quantum-dhcp-agent. This agent was probably restarted several time while rebooting the system (this is an install from ubuntu packages). My understanding is that this should not happen. This may mess up the logic on the nova side for poking iptables holes to allow VMs to receive DHCP, or it may be harmless, I'm not sure.

root@openstack-cc:~# quantum port-list -c device_owner -c fixed_ips
+--------------------------+------------------------------------------------------------------------------------+
| device_owner | fixed_ips |
+--------------------------+------------------------------------------------------------------------------------+
| network:dhcp | {"subnet_id": "f9bbbb9b-5d93-4228-9d3b-0c58f3ba2fb6", "ip_address": "192.168.0.2"} |
| network:router_interface | {"subnet_id": "c37d9f81-eede-4ac6-9397-d10c240a3d11", "ip_address": "60.0.0.1"} |
| compute:nova | {"subnet_id": "c37d9f81-eede-4ac6-9397-d10c240a3d11", "ip_address": "60.0.0.3"} |
| network:router_gateway | {"subnet_id": "f9bbbb9b-5d93-4228-9d3b-0c58f3ba2fb6", "ip_address": "192.168.0.4"} |
| network:dhcp | {"subnet_id": "c37d9f81-eede-4ac6-9397-d10c240a3d11", "ip_address": "60.0.0.2"} |
| network:dhcp | {"subnet_id": "c37d9f81-eede-4ac6-9397-d10c240a3d11", "ip_address": "60.0.0.4"} |
| compute:nova | {"subnet_id": "bb474b8f-df9c-444c-8782-8bcf7bacbd1c", "ip_address": "61.0.0.3"} |
| network:dhcp | {"subnet_id": "bb474b8f-df9c-444c-8782-8bcf7bacbd1c", "ip_address": "61.0.0.2"} |
| compute:nova | {"subnet_id": "f9bbbb9b-5d93-4228-9d3b-0c58f3ba2fb6", "ip_address": "192.168.0.5"} |
| network:router_interface | {"subnet_id": "bb474b8f-df9c-444c-8782-8bcf7bacbd1c", "ip_address": "61.0.0.1"} |
| network:router_gateway | {"subnet_id": "f9bbbb9b-5d93-4228-9d3b-0c58f3ba2fb6", "ip_address": "192.168.0.3"} |
+--------------------------+------------------------------------------------------------------------------------+

Revision history for this message
dan wendlandt (danwent) wrote :

mark, could you comment here? Thanks.

Revision history for this message
yong sheng gong (gongysh) wrote :

Dan, can u print the device_id?
It seems u have run the dhcp-agent in diffrent host? or run many instances of dhcp-agent in multi hosts.

Revision history for this message
dan wendlandt (danwent) wrote :

here is is with device_Id.

| dhcp80a81c5c-6cb9-5124-b268-7d352d654d3c-68fbbf41-ff2f-4bf3-a850-369518cb6d3b | network:dhcp | {"subnet_id": "c37d9f81-eede-4ac6-9397-d10c240a3d11", "ip_address": "60.0.0.2"} |
| dhcpfba529fc-becd-5f63-b6f3-2f92fed6a2b7-68fbbf41-ff2f-4bf3-a850-369518cb6d3b | network:dhcp | {"subnet_id": "c37d9f81-eede-4ac6-9397-d10c240a3d11", "ip_address": "60.0.0.4"} |

Looking at the code, I see that device_id is generated based on DNS name. I changed the DNS name of the host after running the dhcp agent, so I guess that explains it. I'll close this bug. I was really just making sure there was not a serious bug here, and the change of DNS name seems to explain it.

Changed in quantum:
status: New → Won't Fix
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.