... because database options of the base trove class have been
deprecated.
Remove CentOS Stream 8 jobs
RDO Zed (currently master) will only support
CentOS Stream 9 [1].
The CS8 DLRN builder is about to be stopped
and removed, so we can get rid of the CS8 jobs.
Plus we need to require "six" explicitly.
Until now, it was installed as transitive dependency.
[1] https://lists.rdoproject.org/archives/list/dev@lists.rdoproject.org/thread/EKPN4RMQEKYS3ZAHRUHZQ47IF7QXOLZH/
Change-Id: Ib82083f328c3e23b07559652e31acb0153c61a43
CentOS Stream 9 based nodes have been included in nodepool
configuration. This patch is adding centos9 jobs in check pipeline for
packstack project.
It's also implementing some changes required for CS9:
- Setting mkisofs as mkisofs_cmd which is alias for genisoimage in cs8
and to xorriso in cs9.
- It enable repo crb and remove redhat-lsb-core package which is not
longer included inc cs9.
- We are forcing to run puppet with LANG=C.UTF-8. In CS9, LANG variable
is not longer being forwarded in ssh connection so we need to force
viable UTF-8 character set anc C.UTF8 is always available [1].
Until we have a proper promotion pipeline for centos9 I'll keep the jobs
as non-voting and only in check pipeline.
[1] https://bugzilla.redhat.com/show_bug.cgi?id=2002734
Change-Id: I32798900e12dd866138ac352334be3ea82deb149
CentOS is publishing some preview composes of CentOS Stream 9. In RDO,
we are working to get ready for it ASAP.
This patch is proposing some changes needed for current status of
CentOS9 so that i can use this patch for it:
- Set rsync to be executed without xinetd (xinetd is not in CS9)
- Added the creation of initiatorname.iscsi using openstacklib::iscsid
It's also creating packstack-centos9 and packstack-multinode-centos9
base jobs so that we can start running packstack jobs in third-party
zuul servers.
Change-Id: I64edd1876ccc32f98ab2330024d9ab513f42a158
The class was deprecated and now removed with [1],
this patch cleans it's usage and use recommended
class to set notification driver.
[1] https://review.opendev.org/#/c/737492/
Also remove epel repos as packages from it
conflicts with RDO repos.
Change-Id: Ie4432221bdb795dd03f1477609b792ae8400a5fc
- Add jobs removed in [1] to the local configuration.
- Fix SSH key distribution in multinode job. It was broken on a
recent Zuul update, it seems.
- Update puppetlabs-apache version to 5.4.0
[1] - https://review.opendev.org/730813
Closes-Bug: #1880681
Depends-On: https://review.opendev.org/730813
Change-Id: I2572d92cbaabc1f02fc0790896ccaa3edfc5e6c2
As part of adoption of CentOS 8 in RDO, this patch is adding support to
test with it:
- Adapt run_tests.sh for CentOS8.
- Add the option to have different versions of puppet modules (needed
for puppet-mysql until we can update mariadb in CentOS7.
- Add centos8 playbooks and jobs definition.
Change-Id: Ie53a8cb5517651487436e24bf60f51bc531f16d6
Fedora 28 is end-of-live and gets removed from OpenDev testing,
remove it completely here. Packstack plans to use CentOS 8 instead of
Fedora for Python 3 based distros.
This will be backported to all open branches.
Change-Id: I10f444de465437db31efd438b07ca18497580d82
Previously, the facter detection for virtual machines could fail
to detect an OpenStack VM as such, creating issues in CI jobs because
we selected KVM virtualization instead of QEMU.
This can be fixed by adding the virt-what binary, which is used by
facter to detect if a machine is virtual or not [1].
[1] - 2aa2d1cd64/lib/src/facts/linux/virtualization_resolver.cc (L62-L64)
Change-Id: I9278887bfdbca209e7f3498273b85edd8efd987f
This patch is creating a new job running on fedora-28 images which
runs scenario-py3.sh to validate python3 builds on fedora.
Job is non-voting.
Additionally this patch is fixing puppetfile to get newer versions,
aligned with what we have in p-o-i.
Change-Id: I404fc6d38dd5177a3aaef51b0450ad2850595e32
Packstack has traditionally allowed multinode deployments with
1 controller, one or more network nodes and several compute nodes,
but this has never been properly tested in our CI.
This job enables an option to use a 2-node architecture, with one
controller and one compute node. It adds a job running scenario002,
but enables scenario001 and scenario003, so they can be used on
demand if needed.
Change-Id: Ide96a232c902370f3ec25ed54f2b272cc1757dad
* Move direct playbook implementation to a independent role so that the role
is reusable in a different playbooks.
* Add a new scenario which installs openstack allinone without running
tempest tests.
Change-Id: I6f19d2b8128fbb0f2034892a11127bce73d102f9
- Do not generate the SSH private key if it already exists
- Set skipsdist = True in tox.ini
- Add Zuul v3 layout
Change-Id: I4b0254bf175d51c0a468b01f8ac034db43859572