yahoo-eng-team team mailing list archive

[gouzongmei <gouzongmei@xxxxxxxxxx>]

Public bug reported:

Environment:
1. Two compute nodes A and B, each of them has four pci devices of "vendor_id":"8086","product_id":"150e", which have all been configured to passthrough_whitelist.
2. Controller nova conf configured to "pci_alias={"vendor_id":"8086", "product_id":"150e", "name":"a1"}"
3. Extra_specs of flavor pci_flavor configured to "{u'pci_passthrough:alias': u'a1:2'}".

Test Steps:
1. Create instance vm1 with pci_flavor, then vm1 is created on A, two of the pci devices were allocated.
2. Migrate vm1 from A to B, then vm1 state changed to error and two pci devices' status on node B changed to "claimed" while two pci devices on A are still "allocated".

** 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/1308517

Title:
  migrating a vm with pci devices  caused DB inconsistent and vm state
  error

Status in OpenStack Compute (Nova):
  New

Bug description:
  Environment:
  1. Two compute nodes A and B, each of them has four pci devices of "vendor_id":"8086","product_id":"150e", which have all been configured to passthrough_whitelist.
  2. Controller nova conf configured to "pci_alias={"vendor_id":"8086", "product_id":"150e", "name":"a1"}"
  3. Extra_specs of flavor pci_flavor configured to "{u'pci_passthrough:alias': u'a1:2'}".

  Test Steps:
  1. Create instance vm1 with pci_flavor, then vm1 is created on A, two of the pci devices were allocated.
  2. Migrate vm1 from A to B, then vm1 state changed to error and two pci devices' status on node B changed to "claimed" while two pci devices on A are still "allocated".

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1308517/+subscriptions