[Bug 680248] Re: Random VLAN subinterfacess fail to come up

[Steve Langasek]

Network startup handling had been significantly reworked for the
upcoming 12.04LTS, and this includes fixes to make vlan starting
reliable. The changes have also been uploaded as an sru to 11.10, which
is still in progress. Please consider testing with oneiric or precise
and feel free to reopen this report if you are still seeing problems.

** Package changed: upstart (Ubuntu) => vlan (Ubuntu)

** Changed in: vlan (Ubuntu)
       Status: New => Fix Released

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to upstart in Ubuntu.
https://bugs.launchpad.net/bugs/680248

Title:
  Random VLAN subinterfacess fail to come up

Status in “vlan” package in Ubuntu:
  Fix Released

Bug description:
  Binary package hint: upstart

  On machines with a fairly large number of VLAN subinterfaces (most of
  our machines have ~20), a random subset of the configured
  subinterfaces sometimes fail to come up on boot. Sometimes all
  interfaces come up correctly, sometimes one or a few subint(s) doesn't
  come up. Whether any subints fail to come up, how many subints fail,
  and which subints appears completely random.

  Verbose output from init(8) is attached. In this example, five
  subints, bond0.47 through bond0.51, failed to come UP/RUNNING despite
  upstart declaring that they entered the 'running' state:

      network-interface (bond0.47) state changed from post-start to
  running

  status(8) shows the interface in start/running, but ifconfig(8) shows
  the subint isn't UP and RUNNING. For example:

      [jwm at virt03.ekgv.ca:pts/0 ~> ifconfig bond0.47
      bond0.47  Link encap:Ethernet  HWaddr e4:1f:13:7b:fa:92  
                BROADCAST MASTER MULTICAST  MTU:1500  Metric:1
                RX packets:0 errors:0 dropped:0 overruns:0 frame:0
                TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
                collisions:0 txqueuelen:0 
                RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B) 
     
      [jwm at virt03.ekgv.ca:pts/0 ~> sudo status network-interface INTERFACE=bond0.47
      network-interface (bond0.47) start/running

  Subints that fail to come up are still listed in
  /var/run/network/ifstate:

      [jwm at virt03.ekgv.ca:pts/0 ~> fgrep bond0.47 /var/run/network/ifstate 
      bond0.47=bond0.47

  Representative sections of /etc/network/interfaces:

      auto bond0
      iface bond0 inet manual
          bond-slaves none

      auto eth0
      iface eth0 inet manual
          bond-master bond0
          bond-primary eth0 eth1

      auto eth1
      iface eth1 inet manual
          bond-master bond0
          bond-primary eth0 eth1

      auto bond0.47
      iface bond0.47 inet manual

      auto br10
      iface br10 inet manual
          bridge_ports bond0.47
          bridge_maxwait 0
      [more subints and bridge interfaces...]

  
  Finally, I'm not sure it's relevant to this behavior, but I thought I'd mention it: these VLAN subinterfaces are part of bridge interfaces. All of our bridge interfaces contain only a single interface (the VLAN subint) at boot; KVM VMs are eventually added to these bridge groups.

  When a VLAN subint fails to come up, so does the corresponding bridge
  interface:

      [jwm at virt03.ekgv.ca:pts/0 ~> sudo status network-interface INTERFACE=br10
      status: Unknown instance: br10
      [jwm at virt03.ekgv.ca:pts/0 ~> ifconfig br10
      br10: error fetching interface information: Device not found
      [jwm at virt03.ekgv.ca:pts/0 ~> fgrep br10 /var/run/network/ifstate
      br10=br10

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/vlan/+bug/680248/+subscriptions