yahoo-eng-team team mailing list archive

[Brian Haley <1696893@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

Since arping is spawned in a thread, if a router is added and quickly
removed from a network, the arping calls could generate errors on the
second or third loop, for example:

Exit code: 2; Stdin: ; Stdout: ; Stderr: arping: Device qr-1e77796c-2b
not available.

This can happen in this scenario:

T(0): internal_network_added()
        port plugged
        arping started in thread

T(1): internal_network_removed()
        port unplugged

T(2): arping fails
T(3): arping fails

An example is in:

http://logs.openstack.org/02/469602/6/check/gate-tempest-dsvm-neutron-
linuxbridge-ubuntu-
xenial/7a048d9/logs/screen-q-l3.txt.gz#_Jun_09_00_23_55_483118

Just search for qr-1e77796c-2b in the logs before this time.

The arping code should detect this on a failure, log a warning and
return early as there is no way to stop the thread once it is spawned.

** Affects: neutron
     Importance: Low
     Assignee: Brian Haley (brian-haley)
         Status: In Progress

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1696893

Title:
  Arping code should detect missing interface and return early

Status in neutron:
  In Progress

Bug description:
  Since arping is spawned in a thread, if a router is added and quickly
  removed from a network, the arping calls could generate errors on the
  second or third loop, for example:

  Exit code: 2; Stdin: ; Stdout: ; Stderr: arping: Device qr-1e77796c-2b
  not available.

  This can happen in this scenario:

  T(0): internal_network_added()
          port plugged
          arping started in thread

  T(1): internal_network_removed()
          port unplugged

  T(2): arping fails
  T(3): arping fails

  An example is in:

  http://logs.openstack.org/02/469602/6/check/gate-tempest-dsvm-neutron-
  linuxbridge-ubuntu-
  xenial/7a048d9/logs/screen-q-l3.txt.gz#_Jun_09_00_23_55_483118

  Just search for qr-1e77796c-2b in the logs before this time.

  The arping code should detect this on a failure, log a warning and
  return early as there is no way to stop the thread once it is spawned.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1696893/+subscriptions