promotion is failing on tempest.scenario.test_network_basic_ops.TestNetworkBasicOps

[Emilien Macchi]

Promotion jobs fail on http://logs.openstack.org/periodic/periodic-tripleo-ci-centos-7-scenario004-multinode-oooq/ce0102f/logs/undercloud/home/jenkins/tempest_output.log.txt.gz#_2017-09-10_07_51_47

2017-09-10 07:51:47 | ==============================
2017-09-10 07:51:47 | Failed 1 tests - output below:
2017-09-10 07:51:47 | ==============================
2017-09-10 07:51:47 |
2017-09-10 07:51:47 | setUpClass (tempest.scenario.test_network_basic_ops.TestNetworkBasicOps)
2017-09-10 07:51:47 | ------------------------------------------------------------------------
2017-09-10 07:51:47 |
2017-09-10 07:51:47 | Captured traceback:
2017-09-10 07:51:47 | ~~~~~~~~~~~~~~~~~~~
2017-09-10 07:51:47 | Traceback (most recent call last):
2017-09-10 07:51:47 | File "/usr/lib/python2.7/site-packages/tempest/test.py", line 251, in setUpClass
2017-09-10 07:51:47 | six.reraise(etype, value, trace)
2017-09-10 07:51:47 | File "/usr/lib/python2.7/site-packages/tempest/test.py", line 241, in setUpClass
2017-09-10 07:51:47 | cls.setup_clients()
2017-09-10 07:51:47 | File "/usr/lib/python2.7/site-packages/tempest/scenario/manager.py", line 82, in setup_clients
2017-09-10 07:51:47 | cls.volumes_client = cls.os_primary.volumes_v2_client
2017-09-10 07:51:47 | AttributeError: 'Manager' object has no attribute 'volumes_v2_client'

[Alan Pevec (apevec)]

Previous occurrence of this error was resolved after Tempest was pinned to 17.0.0 https://bugs.launchpad.net/tripleo/+bug/1714577

This pin is still the same, what changed?

[Alan Pevec (apevec)]

Looking at http://logs.openstack.org/periodic/periodic-tripleo-ci-centos-7-scenario004-multinode-oooq/ce0102f/logs/undercloud/var/log/yum.log.txt.gz

Installed: 1:openstack-tempest-16.1.0-2.e70e0fegit.el7.noarch

This is CBS build from pike-testing, which should not be installed, master trunk repo has latest Tempest 17:

https://trunk.rdoproject.org/centos7-master/current-tripleo/openstack-tempest-17.0.1-0.20170908221615.fc28933.el7.centos.noarch.rpm

Looks like yum priorities are mixed up:

http://logs.openstack.org/periodic/periodic-tripleo-ci-centos-7-scenario004-multinode-oooq/ce0102f/logs/undercloud/etc/yum.repos.d/delorean.repo.txt.gz
priority=20

http://logs.openstack.org/periodic/periodic-tripleo-ci-centos-7-scenario004-multinode-oooq/ce0102f/logs/undercloud/etc/yum.repos.d/delorean-pike-testing.repo.txt.gz
priority=10

so deps repo (pike-testing) is taking precedence, was this changed recently by accident in oooq repo setup?

[Alan Pevec (apevec)]

Looks like priority=10 for deps was incorrectly set in https://review.openstack.org/502102

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/502360

[Alan Pevec (apevec)]

YUM priorities fix https://review.openstack.org/502360

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/502360
Committed: https://git.openstack.org/cgit/openstack/tripleo-quickstart/commit/?id=5c2f6d206284dfdd3a8c37fcc90941590b93ee1f
Submitter: Jenkins
Branch: master

commit 5c2f6d206284dfdd3a8c37fcc90941590b93ee1f
Author: Alan Pevec <email address hidden>
Date: Mon Sep 11 00:32:34 2017 +0200

    Set correct priority for deps repo

    This was incorrectly set in https://review.openstack.org/502102
    yum priorities are:
    "current" 10 - highest prio
    "current-tripleo" 20
    deps (RELEASE-testing) 30 - lowest prio

    Change-Id: I00f3f2294e912a7a920c65085601f8faa2bb1a96
    Partial-Bug: #1716239