Race conditions with builds and deletes in Ironic driver

[Chris Behrens]

While perhaps not officially supported by nova, when running 2 nova-computes for Ironic, you can get a delete to happen at the same time as a build. The ironic virt driver for nova skips unprovision when the node is not in certain states... It happens that unprovision is skipped when a Node is still deploying... Nova then tries to unset the instance_uuid and ends up failing and retrying on a 409 until ironic finishes the build. It then succeeds and you end up with a Node in ironic that is 'active' but has no instance_uuid.

In addition to this, there's no provision_state checking in the virt driver when reporting resources to the compute manager. There's checking on if instance_uuid is assigned.. and if there's not one, it assumes the node is free. In this case, it turns out the node is not really free... but somewhat orphaned. Scheduling should probably skip these if they happen.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/104649

[Dmitry Tantsur (divius)]

Hi Chris, could you give status update on this bug? The patch seems abandoned and also should be moved to Nova.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Lucas Alvares Gomes (<email address hidden>) on branch: master
Review: https://review.openstack.org/104649
Reason: Thanks for the patch! Unfortunately the changes doesn't belongs to Ironic anymore, the Ironic Nova driver was finally moved to the Nova tree (yay). Please resubmit it to Nova.

Thanks again!

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/135767

[OpenStack Infra (hudson-openstack)]

Change abandoned by Sean Dague (<email address hidden>) on branch: master
Review: https://review.openstack.org/135767
Reason: This review is > 4 weeks without comment, and failed Jenkins the last time it was checked. We are abandoning this for now. Feel free to reactivate the review by pressing the restore button and leaving a 'recheck' comment to get fresh test results.

[Davanum Srinivas (DIMS) (dims-v)]

Removing "In Progress" status and assignee as change is abandoned.

[Jim Rollenhagen (jim-rollenhagen)]

we should still push this through, even though it's not fully supported. I'm happy to pick this one up.

[Michael Davies (mrda)]

Assigned to Jim as per his lying under the bus

[Markus Zoeller (markus_z) (mzoeller)]

This is an automated cleanup. This bug report has been closed because it
is older than 18 months and there is no open code change to fix this.
After this time it is unlikely that the circumstances which lead to
the observed issue can be reproduced.

If you can reproduce the bug, please:
* reopen the bug report (set to status "New")
* AND add the detailed steps to reproduce the issue (if applicable)
* AND leave a comment "CONFIRMED FOR: <RELEASE_NAME>"
  Only still supported release names are valid (LIBERTY, MITAKA, OCATA, NEWTON).
  Valid example: CONFIRMED FOR: LIBERTY