"nova host-servers-migrate <host>" Migrate instances on free compute in error status

[Egor Kotko]

{"build_id": "2014-01-29_13-30-05", "ostf_sha": "338ddf840c229918d1df8c6597588b853d02de4c", "build_number": "67", "nailgun_sha": "3463912a986465133058a24c615c3548cef53cac", "fuelmain_sha": "7d8768f2ac7e1e54d16c135e4ebd64722e49179e", "astute_sha": "200f68381327d955428c371582c03a97bfec3154", "release": "4.1", "fuellib_sha": "73e74f0c449ad86b3da922c8bd5eb333eac94489"}

"nova host-servers-migrate <host>" Migrate instances on free compute in error status

Steps to reproduce
1. Get ISO#67
2. Cluster configuration: Ubuntu, simple, 1Controller, (2+ceph)Computes, Ceph for images, Neutron GRE.
3. Create Instance check on wich compute it is.
4. On controller execute command "nova --debug host-servers-migrate <host>" host should be the name of compute with created instance.

Expected result:
Instance should migrate and be in Active status

Actual result:
After migration instance became in Error state.

compute node on wich should be migrated this instance contain following:

<180>Jan 29 16:12:16 node-3 nova-nova.compute.manager AUDIT: Migrating
<182>Jan 29 16:12:18 node-3 nova-nova.virt.libvirt.driver INFO: Creating image
<179>Jan 29 16:12:19 node-3 nova-nova.virt.libvirt.imagebackend ERROR: error opening rbd image /var/lib/nova/instances/_base/4e3fb6726c8ee72072724a16179d5e400c712864
Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/nova/virt/libvirt/imagebackend.py", line 467, in __init__
    read_only=read_only)
  File "/usr/lib/python2.7/dist-packages/rbd.py", line 351, in __init__
    raise make_ex(ret, 'error opening image %s at snapshot %s' % (name, snapshot))
ImageNotFound: error opening image /var/lib/nova/instances/_base/4e3fb6726c8ee72072724a16179d5e400c712864 at snapshot None
<179>Jan 29 16:12:19 node-3 nova-nova.compute.manager ERROR: Setting instance vm_state to ERROR
Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 3153, in finish_resize
    disk_info, image)
  File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 3121, in _finish_resize
    block_device_info, power_on)

[Xav Paice (xavpaice)]

This can also be reproduced by:

* Havana 2013.2.2, 3 ceph/compute nodes, 1 controller. Glance uses RBD for image store, so does Cinder.
* no shared storage under /var/lib/nova/instances
* create an instance from an image
* nova migrate <instance>
* resulting in the instance being in 'error' state, and the same exception as noted in the original report.

Note that nova live-migration works fine, but we noted the problem when attempting to resize an instance.

The image listed in the exception under var/lib/nova/instances/_base/ does not appear to be the uuid of any image listed by nova image-list.

[Dmitry Borodaenko (angdraug)]

I can't reproduce this bug on Icehouse version of Nova patched with this patch series:
https://github.com/angdraug/nova/tree/rbd-ephemeral-clone-stable-icehouse

I suspect that it's yet another bug fixed by:
https://review.openstack.org/91722

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

Per #3, this no longer seems to be the problem. please reopen if necessary