Wrong synchronized processes of creating resources in pacemaker

[Gleb]

err: /Stage[main]/Galera/Service[mysql]: Could not evaluate: resource p_mysql not found

Yesterday this bug reproduced on hardware environment. Sergey Vasilenko can provide the detail explanation and solutions.
I just filed this bug to not lost this information.

[Sergey Vasilenko (xenolog)]

We need refactor some parts of corosync resources from sync to async approach.

Also may be good idea change crmsh cli utility to pcs.

[Sergey Vasilenko (xenolog)]

This bug happens only on high productivity servers and can't couch on virtual environments.

[Bogdan Dobrelya (bogdando)]

investigate block_until_ready usage for corresponding cs_commit, cs_shadow, etc puppet providers for corosycnc - it could be a source of the sync issues

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/72911

[Bogdan Dobrelya (bogdando)]

exists? method should be provided as well to make the results of commit and shadow providers ensureable

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/72911
Committed: https://git.openstack.org/cgit/stackforge/fuel-library/commit/?id=c309af46a485cc850b77d7c5422bbaa7e3bd5bb5
Submitter: Jenkins
Branch: master

commit c309af46a485cc850b77d7c5422bbaa7e3bd5bb5
Author: Bogdan Dobrelya <email address hidden>
Date: Wed Feb 12 12:10:24 2014 +0200

    Add missed block_until_ready calls for corosync provider

    Closes-bug: #1279211

    Change-Id: I995d01bd28ea8ef41f752feb252be49d19b62c83
    Signed-off-by: Bogdan Dobrelya <email address hidden>