Missing file on slave after ceph migration

[Mikhail Samoylov]

Steps for reproduce:
1. Create cluster
        2. Add 1 node with controller and ceph OSD roles
        3. Add 2 nodes with compute and ceph OSD roles
        4. Deploy the cluster
        5. Check ceph status
        6. Run OSTF
        7. Create a new VM, assign floating ip
        8. Migrate VM
        9. Check cluster and server state after migration
        10. Terminate VM
        11. Check that DHCP lease is not offered for MAC of deleted VM
        12. Create a new VM for migration, assign floating ip
        13. Create a volume and attach it to the VM
        14. Create filesystem on the new volume and mount it to the VM
        15. Migrate VM
        16. Mount the volume after migration
        17. Check cluster and server state after migration
        18. Terminate VM

Actual result:
Step 16 failed with error:
AssertionError: File is absent in /mnt because in mnt we haven't file "file-on-volume"
http://paste.openstack.org/show/507659/

Expected result: Test passed

ISO: 432 MOS
Test code:
https://github.com/openstack/fuel-qa/blob/master/fuelweb_test/tests/test_ceph.py#L542

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/326013

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/326013
Committed: https://git.openstack.org/cgit/openstack/fuel-qa/commit/?id=d28b6868f56ab6f8e40bdbfd8e5c433c17be0f9e
Submitter: Jenkins
Branch: master

commit d28b6868f56ab6f8e40bdbfd8e5c433c17be0f9e
Author: Maksim Malchuk <email address hidden>
Date: Mon Jun 6 19:00:04 2016 +0300

    Fix instance volume mount script creation

    Now the volume mount script would be created with correct newlines.
    Also, this change contain some additional debugging information, so
    the process of creating and mounting the volume logged to the
    instance /var/log/messages through syslog. Also, this commit contain
    the the workaround for the AuthenticationException because the
    Cirros image starts ssh servece just before the cloud-init and in
    some cases we need to wait when the instance_initial_scenario canges
    the password for the cirros user.

    Change-Id: Icb981c90a5ed8beb75fa119158161bf5ea3e5900
    Closes-Bug: #1588808
    Signed-off-by: Maksim Malchuk <email address hidden>

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/mitaka
Review: https://review.openstack.org/326489

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/326489
Committed: https://git.openstack.org/cgit/openstack/fuel-qa/commit/?id=235edd3495efdf93dd11c3c8d683bc15ac634990
Submitter: Jenkins
Branch: stable/mitaka

commit 235edd3495efdf93dd11c3c8d683bc15ac634990
Author: Maksim Malchuk <email address hidden>
Date: Mon Jun 6 19:00:04 2016 +0300

    Fix instance volume mount script creation

    Now the volume mount script would be created with correct newlines.
    Also, this change contain some additional debugging information, so
    the process of creating and mounting the volume logged to the
    instance /var/log/messages through syslog. Also, this commit contain
    the the workaround for the AuthenticationException because the
    Cirros image starts ssh servece just before the cloud-init and in
    some cases we need to wait when the instance_initial_scenario canges
    the password for the cirros user.

    Change-Id: Icb981c90a5ed8beb75fa119158161bf5ea3e5900
    Closes-Bug: #1588808
    Signed-off-by: Maksim Malchuk <email address hidden>
    (cherry picked from commit d28b6868f56ab6f8e40bdbfd8e5c433c17be0f9e)

[Andrey Sledzinskiy (asledzinskiy)]

Test passed on latest swarm run https://product-ci.infra.mirantis.net/job/9.0.system_test.ubuntu.ceph_ha_one_controller/134/