Instance remains in 'resizing' state if network node is down during resize

[Unmesh Gurjar]

Scenario: While an instance is running (Active state), stop the network service and then resize the instance.

Expected Behavior: Instance resize should fail and instance vm_state should change to 'error'.

Actual Behavior: Instance resizing fails, however, instance remains in vm_state=resizing, task_state=resize_prep, power_state=1.

Branch: master

The instance remains in this state even after the network service comes up again. The instance (in resizing state) cannot be deleted/rebooted.

There needs to be a periodic task that monitors such instances (being in resizing state for long time) and set their vm_state to error.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/6632

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/6632
Committed: http://github.com/openstack/nova/commit/4eafbbeb6545dddedb88f8e3d14c7dc13bbf24c5
Submitter: Jenkins
Branch: master

commit 4eafbbeb6545dddedb88f8e3d14c7dc13bbf24c5
Author: Unmesh Gurjar <email address hidden>
Date: Wed Apr 18 11:10:02 2012 +0530

    Updated instance state on resize error.

    Fixes bug 980685.
    Set instance vm_state to ERROR if network node down during resize.

    Change-Id: I2e2f7a98be4b2edfc7d51a4d8d9e31b7b418f2ba