nova-compute needs restarting on AIO deployed w/ ceph

[Matt Thompson]

Deploying a new AIO w/ ceph support enabled always results in nova-compute needing to be restarted.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/214557

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/214557
Committed: https://git.openstack.org/cgit/stackforge/os-ansible-deployment/commit/?id=51757afe1625f6ed74b3bae56f9864c0ea0f6ec8
Submitter: Jenkins
Branch: master

commit 51757afe1625f6ed74b3bae56f9864c0ea0f6ec8
Author: Matt Thompson <email address hidden>
Date: Wed Aug 19 11:41:42 2015 +0100

    Trigger restart after adding user to cephkeys

    Currently, on an AIO, when you deploy w/ ceph support nova-compute
    comes up non-functional and needs a restart. Doing an strace on the
    nova-compute process, you can see nova-compute is unable to read the
    /etc/ceph/ceph.client.cinder.keyring file. This reason for this is due
    to the fact that cinder-volume and nova-compute both live on the same
    host, and a bulk of the tasks in ceph_auth.yml (which do have valid
    restarts) are being taken care of when the cinder installation playbook
    is run. When the nova installation playbook is run, ceph_auth.yml is
    included after nova is installed but when nova is added to cephkeys
    there is nothing else happening in that task file to trigger a restart
    of nova.

    Change-Id: I61e2709ce48ff80fd6eded9c1e6bd649bcd73088
    Closes-Bug: #1486504

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/216320
Committed: https://git.openstack.org/cgit/stackforge/os-ansible-deployment/commit/?id=fd9aa8baaddc0bbc7b9d4ad711bafa8aedc6f0bb
Submitter: Jenkins
Branch: kilo

commit fd9aa8baaddc0bbc7b9d4ad711bafa8aedc6f0bb
Author: Matt Thompson <email address hidden>
Date: Wed Aug 19 11:41:42 2015 +0100

    Trigger restart after adding user to cephkeys

    Currently, on an AIO, when you deploy w/ ceph support nova-compute
    comes up non-functional and needs a restart. Doing an strace on the
    nova-compute process, you can see nova-compute is unable to read the
    /etc/ceph/ceph.client.cinder.keyring file. This reason for this is due
    to the fact that cinder-volume and nova-compute both live on the same
    host, and a bulk of the tasks in ceph_auth.yml (which do have valid
    restarts) are being taken care of when the cinder installation playbook
    is run. When the nova installation playbook is run, ceph_auth.yml is
    included after nova is installed but when nova is added to cephkeys
    there is nothing else happening in that task file to trigger a restart
    of nova.

    Change-Id: I61e2709ce48ff80fd6eded9c1e6bd649bcd73088
    Closes-Bug: #1486504
    (cherry picked from commit 51757afe1625f6ed74b3bae56f9864c0ea0f6ec8)

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

This issue was fixed in the openstack/openstack-ansible 11.2.11 release.

[Doug Hellmann (doug-hellmann)]

This issue was fixed in the openstack/openstack-ansible 11.2.12 release.

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

This issue was fixed in the openstack/openstack-ansible 11.2.14 release.