[library][patching] during patching failed resources are not migrated from the node

[Vladimir Kuklin]

Deploy 5.0 master node

deploy 5.0 env

upgrade 5.1 to latest 5.1, e.g.

{"build_id": "2014-09-04_00-01-17", "ostf_sha": "4dcd99cc4bfa19f52d4b87ed321eb84ff03844da", "build_number": "496", "auth_required": true, "api": "1.0", "nailgun_sha": "3d965a7acd9361a361da92970d5bc5a1be77c969", "production": "docker", "fuelmain_sha": "435c43fe558898566ef8254ec09cc12bcb809c4b", "astute_sha": "8cc2c5b29d75ec2f60787e225564426a0508b3e4", "feature_groups": ["mirantis"], "release": "5.1", "fuellib_sha": "bc23f1359b35ec90d5c7bc9f33778b4f2922d409"}

apply https://review.openstack.org/#/c/118432/ to astute

start patching to 5.0.2
expected behaviour:

pacemaker should move failed resources (e.g. neutron agents) to other nodes

actual result:

resources stay on the node being patched in unmanaged (FAILED) state and puppet run cannot clean them up

[Vladimir Kuklin (vkuklin)]

https://review.openstack.org/#/c/120469/

[OpenStack Infra (hudson-openstack)]

Change abandoned by Vladimir Kuklin (<email address hidden>) on branch: master
Review: https://review.openstack.org/119209

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/5.1
Review: https://review.openstack.org/120479

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/120469
Committed: https://git.openstack.org/cgit/stackforge/fuel-astute/commit/?id=a3e5da62af91b99f958ab958161d3dcec09c657b
Submitter: Jenkins
Branch: master

commit a3e5da62af91b99f958ab958161d3dcec09c657b
Author: Dmitry Ilyin <email address hidden>
Date: Wed Sep 10 20:21:51 2014 +0400

    Add hooks to workaround patching bugs

    Add several pre and post patching hooks and tests
    for them.

    * Stop or migrate corosync services from controller node
      and return them after the deployment

    * Stop openstack services and kill pids that was
      not stopped together with their children

    * Remove openstack packages before deployment
      to prevent conflicts during the update and rollback

    Change-Id: Icfc17308d955e55b5e5b31a9c17e0b6d13d8ba10
    Closes-bug: #1364081
    Closes-bug: #1365635
    Closes-bug: #1364068
    Co-Authored-By: Vladimir Sharshov <email address hidden>
    Co-Authored-By: Bogdan Dobrelya <email address hidden>
    Co-Authored-By: Vladimir Kuklin <email address hidden>

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/120479
Committed: https://git.openstack.org/cgit/stackforge/fuel-astute/commit/?id=f5fbd89d1e0e1f22ef9ab2af26da5ffbfbf24b13
Submitter: Jenkins
Branch: stable/5.1

commit f5fbd89d1e0e1f22ef9ab2af26da5ffbfbf24b13
Author: Dmitry Ilyin <email address hidden>
Date: Wed Sep 10 20:21:51 2014 +0400

    Add hooks to workaround patching bugs

    Add several pre and post patching hooks and tests
    for them.

    * Stop or migrate corosync services from controller node
      and return them after the deployment

    * Stop openstack services and kill pids that was
      not stopped together with their children

    * Remove openstack packages before deployment
      to prevent conflicts during the update and rollback

    Change-Id: Icfc17308d955e55b5e5b31a9c17e0b6d13d8ba10
    Closes-bug: #1364081
    Closes-bug: #1365635
    Closes-bug: #1364068
    Co-Authored-By: Vladimir Sharshov <email address hidden>
    Co-Authored-By: Bogdan Dobrelya <email address hidden>
    Co-Authored-By: Vladimir Kuklin <email address hidden>

[Tatyanka (tatyana-leontovich)]

seems should be marked as invalid according to patching feature should be re-designed