Engine raises the error 'Delayed call not found'

[Nikolay Makhotkin]

Steps to reproduce:

1. Run mistral
2. Run at least 2 engines
3. Run workflow where scheduler will be used (e.g. with policies: wait-before, wait-after or retry)
4. One of engines raises the error 'Delayed call not found'

The reason is that each engine instance also creates the scheduler process. The scheduler check DB on delayed calls every second therefore both schedulers tries to process this delayed call. Once the call was deleted from the DB, another scheduler also tries to delete it but it doesn't find the delayed call because it was already deleted.

The possible solution is to wrap delayed call processing into one transaction.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/161716

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/161716
Committed: https://git.openstack.org/cgit/stackforge/mistral/commit/?id=850e658cbbbdea75817c6a4fb55f64c538631c3b
Submitter: Jenkins
Branch: master

commit 850e658cbbbdea75817c6a4fb55f64c538631c3b
Author: Nikolay Mahotkin <email address hidden>
Date: Thu Mar 5 16:31:28 2015 +0300

    Fixing scheduler work

    Closes-Bug: #1428635

    Change-Id: Ia1c2ddf9cd4e0cbb71121e88fbb25244fae0d80b