XenServer port needs to clear out vm-data/networking before issuing resetnetwork command

Thierry Carrez <email address hidden> wrote:

** Changed in: nova
   Importance: Undecided => Medium

** Changed in: nova
       Status: New => Confirmed

--
You received this bug notification because you are subscribed to
OpenStack.
https://bugs.launchpad.net/bugs/747394

Title:
  XenServer port needs to clear out vm-data/networking before issuing
  resetnetwork command

Status in OpenStack Compute (Nova):
  Confirmed

Bug description:
  The guest agent will pull networking information from vm-
  data/networking in xenstore. If Nova chooses a different name than
  what is already in that directory directory, the guest agent will
  apply all configurations, even if they are old or conflicting.

Confidentiality Notice: This e-mail message (including any attached or
embedded documents) is intended for the exclusive and confidential use of the
individual or entity to which this message is addressed, and unless otherwise
expressly indicated, is confidential and privileged information of Rackspace.
Any dissemination, distribution or copying of the enclosed material is prohibited.
If you receive this transmission in error, please notify us immediately by e-mail
at <email address hidden>, and delete the original message.
Your cooperation is appreciated.

@Johannes: still working on that ? If yes could you refresh your merge proposal with the latest comments and trunk ? If no, please unassign yourself from the bug so that someone else can have a try.

The bug is fixed, but the changes necessary to the test suite are what is holding this up. I haven't had a chance to refactor the code to handle the changes necessary to fix this bug.

I may have some time to work on that this week.

@Johannes: did you have time to refresh that one ?

de-assigning since this is pretty stale now

This is related to:
https://review.openstack.org/#/c/23662/

Sorry, wrong bug

Given the moves towards cloud-init this is much less important, and does seemed to have caused any problems so far.

Hmm, thinking about this, its related to issues with inject_network_info.

Currently people call inject_network_info, and potentially alter xenstore, then call resetnetwork.

So, sadly, this behaviour is now being used as a feature.

It looks like it's fixed rather than invalid.

When VIFs are removed from an instance, the xenstore entry is removed too (this wasn't always the case).

So that effectively fixes the problem, which probably also explains why we haven't seen this problem in a while.