- run `pre-commit autoupdate` and fixed new issues
- adopted newer pre-commit config for ansible-lint 4.2.0+
- fixed some reported broken rules
- temporary disabled few rules, just to contain the size of of review,
planning to drop/fix them in follow-ups.
Change-Id: I807ba4e919527be56c85ec72d0f4c7148f04e994
- Fixes 206 errors and removed the skip.
- Also replaces revision with latest tag (pushed today)
Bug: https://bugs.launchpad.net/tripleo/+bug/1848512
Change-Id: I69d75ba9de8ab0e91eb45fd4f9febfdb28cdcb3a
This patch updates nodepool-setup and multinodes
roles to work with non centos images. It updates following:-
- Add a role variable: "default_image_user" which defaults to
'centos'. It can be used to configure image_user like 'fedora'.
- Use 'package' module instead of 'yum' so it can work with non
yum systems.
- yum_contentdir is not needed for Fedora, so skipped.
- centos-release-openstack and epel repos are not needed for
Fedora, so skipped.
Change-Id: Ia1245a423e608273120b2c6b9668e2126073cccc
Upgrades yamllint to latest version and adots use of its strict
checking.
Fix all known problems reported by yamllint so we don't have to do
that while touching these files.
Change-Id: I4bdc520d9e2aff086c4b463718bc1e053261a4f5
Story: https://tree.taiga.io/project/tripleo-ci-board/task/381
Makes those files conformant with current linting rules and avoids
linting errors when we need to toch them again.
Previous doing "pre-commit run -a" uncovered these errors, now is no
longer reporting any other errors.
Change-Id: Ie4cf229c8f11c2b55b323eac23c89483b26d3781
By making sure we prefix the private net name, we ensure that we can
create multiple reproducer stack on the same cloud provider.
Change-Id: I2fec9abe139481cb1e10bb72aac88b651214269f
Closes-Bug: #1775826
If there are several images with the centos-7-latest tag, the resulting
environment file is broken. This fixes the command call to only return
one result.
Change-Id: If1c339b11e86a25b35c98c1f1e5024d1ee9e42bb
Instead of looping on 'all' nodes, restricting the loop
to only subnodes groups avoid failure if other hosts
is present.
Change-Id: Ife744d89836af2c98bffe5ae10f8544625ed9f2b
Avoids deprecation warnings like below by adopting recommended syntax.
[DEPRECATION WARNING]: Using tests as filters is deprecated. Instead of
using `result|failed` instead use `result is failed`. This feature will
be removed in version 2.9. Deprecation warnings can be disabled by
setting deprecation_warnings=False in ansible.cfg.
Change-Id: Ie0e0fbc61090e2b89971252ee60700158acf5fd4
Signed-off-by: Sorin Sbarnea <ssbarnea@redhat.com>
In order to filter out saved snapshots in the tenant, look for
public images when calculating the last one.
Change-Id: I06099486f9b0be9e7bed40ac6b351658a612b3df
Since the change I9617dafff83f800ed75b4478c7040301b0c944a1 there is
now 2 private network interfaces, so using the networks list to return
subnode_ip_pair now returns 3 entries, which breaks the reproducer.
This change returns the IP addresses from the floating IP, so that any
changes to the subnode networking won't cause issues here again.
Change-Id: I3f20e34e03874cfbd4701c70f9818141568b4d55
Run the new "openstack overcloud failures" to show shows the deployment
failures which will be useful for debug.
Depends-On: I9ce4ab9e6c690abb5aed887f8fe6add4bbf52167
Change-Id: I014a6bed93c62819c423300c166936e44d1e67c1
For standalone reproducing we need another interface (eth1) on the
subnode. Add it to heat template of subnode.
Change-Id: I9617dafff83f800ed75b4478c7040301b0c944a1
There are problems with RDO cloud + multinode communication, there is no
vxlan package arriving from subnode-1 to subnode-0, to allow that vxlan
port needs to be open.
Partial-Bug: 1767099
Change-Id: I9d070e208000194e73654299c9645679bea27169
Using OS::Neutron::Port directly just to associate a floating IP is
strongly discouraged because the scheduling of the server becomes
constrained by the existing port.
Instead, let nova/neutron create the port implicitly and fetch the
resulting port ID to use for the floating IP association.
This change also adds a depends_on:router_interface to ensure the
interface exists before floating IPs attempt to use it.
This may improve the success rate of multinode jobs where nodes go to
ERROR due to schedule failures, but I'm not aware of any specific bug
which is tracking this.
Change-Id: Ie5cdd5a465040a0ece98a6103ad60b4a26610e42