Quantcast
Viewing latest article 2
Browse Latest Browse All 3

Comment by Dave Holland for I've been using OpenStack Juno (RDO, CentOS7) in my lab with Flat network and two compute nodes without any issue. Today I installed Mitaka (RDO, CentOS7) from the scratch and re-created the same flat network setup. Everything seems to work fine inside the OpenStack (instances get response from the DHCP agent and are able to reach the flat network). However OpenStack DHCP agent (DNSMASQ to be precise) answers DHCPINFORM requests from other PCs connected to the same network (that are NOT OpenStack instances) - driving them crazy - see below DNSMASQ log. Why is this happening? I'm pretty sure that I didn't have this problem on my Juno setup. Is there any way to fix this? [root@openstack-se1 ~(keystone_admin)]# neutron subnet-show public_subnet +-------------------+--------------------------------------------------+ | Field | Value | +-------------------+--------------------------------------------------+ | allocation_pools | {"start": "10.4.219.224", "end": "10.4.219.244"} | | cidr | 10.4.0.0/16 | | created_at | 2016-06-09T18:01:12 | | description | | | dns_nameservers | 10.1.1.11 | | enable_dhcp | True | | gateway_ip | 10.4.0.1 | | host_routes | | | id | 74bb197c-8ec0-42d0-b334-e320c50911cc | | ip_version | 4 | | ipv6_address_mode | | | ipv6_ra_mode | | | name | public_subnet | | network_id | c70d07a8-35e2-4415-a8bc-249e7d0404a4 | | subnetpool_id | | | tenant_id | 4f37f5b23a034e89bb71050deae9eca3 | | updated_at | 2016-06-09T18:01:12 | +-------------------+--------------------------------------------------+ tail -f /var/log/neutron/dnsmasq.log Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 DHCPINFORM(tap0a3d2724-eb) 10.4.2.4 d8:cb:8a:42:a3:5e Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 tags: tag0, tap0a3d2724-eb Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 DHCPACK(tap0a3d2724-eb) 10.4.2.4 d8:cb:8a:42:a3:5e IL-ronip-PC Jun 9 21:01:50 dnsmasq-dhcp[11603]: Ignoring duplicate dhcp-option 26 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 1:netmask, 15:domain-name, 3:router, 6:dns-server, Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 44:netbios-ns, 46:netbios-nodetype, 47:netbios-scope, Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 31:router-discovery, 33:static-route, 121:classless-static-route, Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 249, 43:vendor-encap, 252 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 next server: 10.4.219.224 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 1 option: 53 message-type 5 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 54 server-identifier 10.4.219.224 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 1 netmask 255.255.0.0 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 28 broadcast 10.4.255.255 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 14 option: 15 domain-name openstacklocal Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 3 router 10.4.0.1 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 6 dns-server 10.1.1.11 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 2 option: 26 mtu 1400 Jun 9 21:01:53 dnsmasq-dhcp[11603]: 4293967296 available DHCP subnet: 10.4.0.0/255.255.0.0 Jun 9 21:01:53 dnsmasq-dhcp[11603]: 4293967296 tags: tag0, bootp, tap0a3d2724-eb Jun 9 21:01:53 dnsmasq-dhcp[11603]: 4293967296 BOOTP ... (more)

Previous: Answer by alex123 for I've been using OpenStack Juno (RDO, CentOS7) in my lab with Flat network and two compute nodes without any issue. Today I installed Mitaka (RDO, CentOS7) from the scratch and re-created the same flat network setup. Everything seems to work fine inside the OpenStack (instances get response from the DHCP agent and are able to reach the flat network). However OpenStack DHCP agent (DNSMASQ to be precise) answers DHCPINFORM requests from other PCs connected to the same network (that are NOT OpenStack instances) - driving them crazy - see below DNSMASQ log. Why is this happening? I'm pretty sure that I didn't have this problem on my Juno setup. Is there any way to fix this? [root@openstack-se1 ~(keystone_admin)]# neutron subnet-show public_subnet +-------------------+--------------------------------------------------+ | Field | Value | +-------------------+--------------------------------------------------+ | allocation_pools | {"start": "10.4.219.224", "end": "10.4.219.244"} | | cidr | 10.4.0.0/16 | | created_at | 2016-06-09T18:01:12 | | description | | | dns_nameservers | 10.1.1.11 | | enable_dhcp | True | | gateway_ip | 10.4.0.1 | | host_routes | | | id | 74bb197c-8ec0-42d0-b334-e320c50911cc | | ip_version | 4 | | ipv6_address_mode | | | ipv6_ra_mode | | | name | public_subnet | | network_id | c70d07a8-35e2-4415-a8bc-249e7d0404a4 | | subnetpool_id | | | tenant_id | 4f37f5b23a034e89bb71050deae9eca3 | | updated_at | 2016-06-09T18:01:12 | +-------------------+--------------------------------------------------+ tail -f /var/log/neutron/dnsmasq.log Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 DHCPINFORM(tap0a3d2724-eb) 10.4.2.4 d8:cb:8a:42:a3:5e Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 tags: tag0, tap0a3d2724-eb Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 DHCPACK(tap0a3d2724-eb) 10.4.2.4 d8:cb:8a:42:a3:5e IL-ronip-PC Jun 9 21:01:50 dnsmasq-dhcp[11603]: Ignoring duplicate dhcp-option 26 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 1:netmask, 15:domain-name, 3:router, 6:dns-server, Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 44:netbios-ns, 46:netbios-nodetype, 47:netbios-scope, Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 31:router-discovery, 33:static-route, 121:classless-static-route, Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 requested options: 249, 43:vendor-encap, 252 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 next server: 10.4.219.224 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 1 option: 53 message-type 5 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 54 server-identifier 10.4.219.224 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 1 netmask 255.255.0.0 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 28 broadcast 10.4.255.255 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 14 option: 15 domain-name openstacklocal Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 3 router 10.4.0.1 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 4 option: 6 dns-server 10.1.1.11 Jun 9 21:01:50 dnsmasq-dhcp[11603]: 575818415 sent size: 2 option: 26 mtu 1400 Jun 9 21:01:53 dnsmasq-dhcp[11603]: 4293967296 available DHCP subnet: 10.4.0.0/255.255.0.0 Jun 9 21:01:53 dnsmasq-dhcp[11603]: 4293967296 tags: tag0, bootp, tap0a3d2724-eb Jun 9 21:01:53 dnsmasq-dhcp[11603]: 4293967296 BOOTP ... (more)
We had a similar problem with a provider network (VLAN) to Liberty/RHOSP8. PCs on the physical network used an external DHCP provider, but the combination of the Windows DHCP client using DHCPINFORM and OpenStack dnsmasq "leaking" caused similar symptoms. We fixed it by filtering DHCP in the switch.

Viewing latest article 2
Browse Latest Browse All 3

Trending Articles