← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1337214] [NEW] VMware: Fail to boot VM when using VDS or the port gropu be created on different vSwitch

 

Public bug reported:

Fail to boot a instance when using the nova-network, I got the error
message in log file

'InvalidVLANPortGroup: vSwitch which contains the port group
VS5_GUEST_SCODEV_G1_V231 is not associated with the desired physical
adapter. Expected vSwitch is vSwitch0, but the one associated is
vSwitch5.

Currently, the logic in vmware driver is that all ESXi systems must have
the exact same networking configuration (the same PortGroup/vSwitch/pNIC
mapping), which typically isn't the case on customer environments.

In my case, on ESX1 the portgroup could be on vSwitch1, but on ESX2 the portgroup could be on vSwitch2, so one of them would fail.
And also if I use the VDS, it doesn't have a physical adapter associated with it and there is a virtual router/firewall connected to that vSwitch which then acts as the gateway for the different PortGroups on it.

So our vSwitch validation should be enhanced.

** Affects: nova
     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/1337214

Title:
  VMware: Fail to boot VM when using VDS or the port gropu be created on
  different vSwitch

Status in OpenStack Compute (Nova):
  New

Bug description:
  Fail to boot a instance when using the nova-network, I got the error
  message in log file

  'InvalidVLANPortGroup: vSwitch which contains the port group
  VS5_GUEST_SCODEV_G1_V231 is not associated with the desired physical
  adapter. Expected vSwitch is vSwitch0, but the one associated is
  vSwitch5.

  Currently, the logic in vmware driver is that all ESXi systems must
  have the exact same networking configuration (the same
  PortGroup/vSwitch/pNIC mapping), which typically isn't the case on
  customer environments.

  In my case, on ESX1 the portgroup could be on vSwitch1, but on ESX2 the portgroup could be on vSwitch2, so one of them would fail.
  And also if I use the VDS, it doesn't have a physical adapter associated with it and there is a virtual router/firewall connected to that vSwitch which then acts as the gateway for the different PortGroups on it.

  So our vSwitch validation should be enhanced.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1337214/+subscriptions


Follow ups

References