← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1077066] Re: nova-network bridge existing system alias handeling failiure.

 

** Also affects: nova/havana
   Importance: Undecided
       Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1077066

Title:
  nova-network bridge existing system alias handeling failiure.

Status in OpenStack Compute (Nova):
  Fix Released
Status in OpenStack Compute (nova) havana series:
  New

Bug description:
  This:
  https://github.com/openstack/nova/blob/stable/essex/nova/network/linux_net.py#L1033

  Also prevalent in Folsom can cause nova-network to fail with the
  following error:

  ---
  Command: sudo nova-rootwrap ip addr del xxx.xxx.xxx.xxx/24 brd 172.30.255.255 scope global secondary dev br0
  Exit code: 255
  Stdout: ''
  Stderr: 'Error: either "local" is duplicate, or "secondary" is a garbage.\n'
  ---

  This is due to the invalid passing of params to the ip addr del

  irc conversation:

  ---
  Apsu`: oneiroi: Yeah, just verified it on a box, to be sure. Can't specify the "secondary" part in the del call. It's not part of the syntax, because it's not user-controllable. A "secondary" is an IP in the same network as one already on the interface in the same scope
  [4:36pm] Apsu`: Let's see if it's different in Folsom
  [4:37pm] oneiroi: Apsu`: abstracted code and output http://fpaste.org/hFe3/
  [4:37pm] uvirtbot`: New bug: #1077056 in tempest "images whitebox test fails with bad SQL connection URI" [High,Confirmed] https://launchpad.net/bugs/1077056
  [4:37pm] AnilV4 left the chat room. (Read error: Connection reset by peer)
  [4:37pm] Apsu`: Nope. Appears to do the same. https://github.com/openstack/nova/blob/stable/folsom/nova/network/linux_net.py#L1193
  [4:37pm] • Apsu` nods
  [4:37pm] Apsu`: Exactly the issue. It's trying to ensure_bridge, and reorder the IPs so nova's bridge IP is first
  [4:38pm] maurosr left the chat room. (Quit: WeeChat 0.3.8)
  [4:38pm] Apsu`: There's a comment from vish in initialize_gateway_device about the same issue
  [4:39pm] Apsu`: So that field splitting code needs to filter out anything past the scope.
  [4:39pm] Apsu`: It's currently using [1:-1] to skip the first and last
  [4:40pm] oneiroi: indeed, filing a bug will then look at a quick fix, and submit patch
  [4:40pm] Apsu`: Cool.

  ---

  
  Abstracted code showing the issue:

  ---
  #!/usr/bin/env python

  str='''
  6: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP 
      link/ether 00:15:17:aa:46:04 brd ff:ff:ff:ff:ff:ff
      inet aa.bb.cc.dd/16 brd 172.30.255.255 scope global br0
      inet xxx.xxx.xxx.xxx/24 brd 192.168.1.255 scope global br0:1
      inet6 fe80::215:17ff:feaa:4604/64 scope link 
         valid_lft forever preferred_lft forever
  '''

  for line in str.split('\n'):
    fields = line.split()
    if fields and fields[0] == 'inet':
      params = fields[1:-1]
      print params	

  ---

  
  ['aa.bb.cc.dd/16', 'brd', '172.30.255.255', 'scope', 'global']
  ['xxx.xxx.xxx.xxx/24', 'brd', '192.168.1.255', 'scope', 'global']

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1077066/+subscriptions