yahoo-eng-team team mailing list archive

Thread
Date
[Bug 1805824] Re: The dhcp port's address may be messed when the port's network has multiple subnets

To: yahoo-eng-team@xxxxxxxxxxxxxxxxxxx
From: OpenStack Infra <1805824@xxxxxxxxxxxxxxxxxx>
Date: Fri, 14 Dec 2018 06:11:14 -0000
Reply-to: Bug 1805824 <1805824@xxxxxxxxxxxxxxxxxx>
Sender: bounces@xxxxxxxxxxxxx
Reviewed:  https://review.openstack.org/620900
Committed: https://git.openstack.org/cgit/openstack/neutron/commit/?id=9aa7af82210f1ffb6274399648962b65162e17d1
Submitter: Zuul
Branch:    master

commit 9aa7af82210f1ffb6274399648962b65162e17d1
Author: Yang JianFeng <yjf1970231893@xxxxxxxxx>
Date:   Thu Nov 29 20:18:31 2018 +0800

    Fix the bug about DHCP port whose network has multiple subnets.
    
    When a subnet's enable_dhcp attribute is updated, we must restart
    dhcp device. So,when we decide whether 'restart' or
    'reload_allocations' in refresh_dhcp_helper function we only compare
    the cidr of subnets which enabled dhcp.
    
    The previous logic only calls 'restart' when deleting or adding a
    subnet. This may cause the dhcp port not updated when the subnet's
    enable_dhcp is updated to True.
    
    Change-Id: Ic547946ac786c5fab82b4ee7078bf86483f51eb5
    Closes-Bug: #1805824


** Changed in: neutron
       Status: In Progress => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1805824

Title:
  The dhcp port's address may be messed when the port's network has
  multiple subnets

Status in neutron:
  Fix Released

Bug description:
  1、When I update one of the subnets's enable_dhcp attribute from true to false, the dhcp port still has the IP of the subnet I updated.
  Reproducible Steps:
  openstack network create test-dhcp
  openstack subnet create --network test-dhcp --subnet-range 2.2.2.0/24 test-dhcp-subnet01
  openstack subnet create --network test-dhcp --subnet-range 3.3.3.0/24 test-dhcp-subnet02
  openstack subnet set test-dhcp-subnet01 --no-dhcp
  Then I login the host of hold the network's dhcp namespace and check the namesapce. The result is as follows, the IP belong to test-dhcp-subnet01 is still exist.
  ip netns exec qdhcp-7ba959b7-299e-4f6c-a806-b6846e1dde68 ip a
  1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
      link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
      inet 127.0.0.1/8 scope host lo
         valid_lft forever preferred_lft forever
      inet6 ::1/128 scope host
         valid_lft forever preferred_lft forever
  18: tapf258f8cc-0d: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN group default qlen 1000
      link/ether fa:16:3e:9b:3d:7a brd ff:ff:ff:ff:ff:ff
      inet 2.2.2.2/24 brd 2.2.2.255 scope global tapf258f8cc-0d
         valid_lft forever preferred_lft forever
      inet 169.254.169.254/16 brd 169.254.255.255 scope global tapf258f8cc-0d
         valid_lft forever preferred_lft forever
      inet 3.3.3.3/24 brd 3.3.3.255 scope global tapf258f8cc-0d
         valid_lft forever preferred_lft forever
      inet6 fe80::f816:3eff:fe9b:3d7a/64 scope link
         valid_lft forever preferred_lft forever
  Next, I restart the neutron-dhcp-agent and check the namesapce again. The IP belong to test-dhcp-subnet01 is disappeared.
  systemctl restart neutron-dhcp-agent
  ip netns exec qdhcp-7ba959b7-299e-4f6c-a806-b6846e1dde68 ip a
  1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
      link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
      inet 127.0.0.1/8 scope host lo
         valid_lft forever preferred_lft forever
      inet6 ::1/128 scope host
         valid_lft forever preferred_lft forever
  18: tapf258f8cc-0d: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN group default qlen 1000
      link/ether fa:16:3e:9b:3d:7a brd ff:ff:ff:ff:ff:ff
      inet 169.254.169.254/16 brd 169.254.255.255 scope global tapf258f8cc-0d
         valid_lft forever preferred_lft forever
      inet 3.3.3.3/24 brd 3.3.3.255 scope global tapf258f8cc-0d
         valid_lft forever preferred_lft forever
      inet6 fe80::f816:3eff:fe9b:3d7a/64 scope link
         valid_lft forever preferred_lft forever

  2、When I update one of the subnets's enable_dhcp attribute from false to true, the DHCP port's IP belong to the subnet I updated no be set.
  Reproducible Steps:
  openstack network create test-dhcp
  openstack subnet create --network test-dhcp --subnet-range 2.2.2.0/24 test-dhcp-subnet01
  openstack subnet create --network test-dhcp --subnet-range 3.3.3.0/24 test-dhcp-subnet02 --no-dhcp
  openstack subnet set test-dhcp-subnet02 --dhcp
  Then I login the host of hold the network's dhcp namespace and check the namesapce. The result is as follows, the IP belong to test-dhcp-subnet02 is missing:
  ip netns exec qdhcp-08621ab0-e012-459c-9a1c-a8f24b56f881 ip a
  1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
      link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
      inet 127.0.0.1/8 scope host lo
         valid_lft forever preferred_lft forever
      inet6 ::1/128 scope host
         valid_lft forever preferred_lft forever
  19: tape7959c7d-39: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN group default qlen 1000
      link/ether fa:16:3e:a8:4d:fe brd ff:ff:ff:ff:ff:ff
      inet 2.2.2.2/24 brd 2.2.2.255 scope global tape7959c7d-39
         valid_lft forever preferred_lft forever
      inet 169.254.169.254/16 brd 169.254.255.255 scope global tape7959c7d-39
         valid_lft forever preferred_lft forever
      inet6 fe80::f816:3eff:fea8:4dfe/64 scope link
         valid_lft forever preferred_lft forever
  Then I restart the neutron-dhcp-agent, The result becomes correct.
  1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
      link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
      inet 127.0.0.1/8 scope host lo
         valid_lft forever preferred_lft forever
      inet6 ::1/128 scope host
         valid_lft forever preferred_lft forever
  19: tape7959c7d-39: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UNKNOWN group default qlen 1000
      link/ether fa:16:3e:a8:4d:fe brd ff:ff:ff:ff:ff:ff
      inet 2.2.2.2/24 brd 2.2.2.255 scope global tape7959c7d-39
         valid_lft forever preferred_lft forever
      inet 169.254.169.254/16 brd 169.254.255.255 scope global tape7959c7d-39
         valid_lft forever preferred_lft forever
      inet 3.3.3.2/24 brd 3.3.3.255 scope global tape7959c7d-39
         valid_lft forever preferred_lft forever
      inet6 fe80::f816:3eff:fea8:4dfe/64 scope link
         valid_lft forever preferred_lft forever

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1805824/+subscriptions
References

[Bug 1805824] [NEW] The dhcp port's address may be messed when the port's network has multiple subnets
From: yangjianfeng, 2018-11-29