TripleO CI should deploy Ceph packages from OpenStack AFS mirrors

[Emilien Macchi]

OpenStack Infra provides AFS mirrors and it include Ceph packages (in fact, Puppet OpenStack CI is already using it and it works pretty well).

To avoid issues when trying to reach centos.org repos, let's use the mirrors provided by OpenStack Infra.

Note: for OVB jobs, since we don't run the nodes via OpenStack Infra cloud providers, we don't have a mirror for the RH1 region, though we can pick a random mirror close to the RH1 location.

[Emilien Macchi (emilienm)]

Note: we want to avoid this situation: http://logs.openstack.org/27/446927/1/check-tripleo/gate-tripleo-ci-centos-7-ovb-ha/d602f1f/console.html#_2017-03-27_10_18_39_088665

[Paul Belanger (pabelanger)]

Yes, please.

We can update configure_mirror.sh in nodepool to setup a .repo file, but disable it. They, you can check to see if the repo configured, and enable it. Its baseurl will be properly pointed to the right mirror.

Otherwise, you'll have to change the baseurl yourself, which is also fine.

[Emilien Macchi (emilienm)]

https://review.openstack.org/#/c/450466/

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/450466
Committed: https://git.openstack.org/cgit/openstack/tripleo-quickstart/commit/?id=27d4a19d7fb10a3a425f4664d91f55ef364dd927
Submitter: Jenkins
Branch: master

commit 27d4a19d7fb10a3a425f4664d91f55ef364dd927
Author: Sagi Shnaidman <email address hidden>
Date: Tue Mar 28 00:41:54 2017 +0300

    Use AFS mirrors for centos hammer and jewel repos

    When running CI jobs we need to use available AFS mirrors to avoid
    network connectivity problems.
    Close-Bug: #1676421
    Change-Id: I7abdbe2806f0d8eb66edc40ade6f67d75d875c33

[Emilien Macchi (emilienm)]

There are no currently open reviews on this bug, changing the status back to the previous state and unassigning. If there are active reviews related to this bug, please include links in comments.

[Alex Schultz (alex-schultz)]

I believe this has been addressed. If this is still a problem, please reopen this bug.