In trove-taskmanager.conf a typo has been
introduced for nova_proxy_admin_tenant_name
option.
It currently is:
"nova_proxy_admin_tenant_name = services"
But should be:
"nova_proxy_admin_tenant_name = service"
Change-Id: I7b5d0ca4c6c994b6dd3c5de3f0a79637fda88177
Closes-Bug: #1770262
The kolla-ansible certificates command creates a self-signed
certificate and corresponding private key. This change sets the
file mode on the files that contain the private key so that they
are not world readable.
Change-Id: I7bfba5975af8d230e5d00c2ce801a914fbbf2d4e
This commit adds the --by-service flag in the discover_hosts
command of the nova cell subsystem, using this flag,
hosts are mapped based on the existing nova-compute services instead
of using the compute node reference.
In this way, we can ensure that the host mappings exists
for this new host (and create it if not) preventing the
HostMappingNotFound exception to loop over and over.
Bug: #1768094
Change-Id: I948004031bc3e125f929ed3e6c670aea11e73a31
Signed-off-by: Jorge Niedbalski <jorge.niedbalski@linaro.org>
Change I0c1328237d2109b1d1d5dcb40f331fff1ab10257 recently prevented
running kolla deployment CI jobs against changes to some files,
including requirements.txt. This change reinstates testing of changes to
requirements.txt, since this can have an effect on the operation of
kolla ansible.
Change-Id: I3d839de9aed60cddebea7d1140f618d3261cca2a
While fixing test-requirement.txt, I found that
check pipeline can be more optimized in term of selecting
the jobs to run based on file change.
- https://review.openstack.org/#/c/567083/
kolla main jobs are not needed to be run on check
pipeline for test-requirements.txt, releasenotes/* etc.
This commit skip those jobs when only above files
are changed in patch. This is will save time as well
as infra resource also.
Change-Id: I0c1328237d2109b1d1d5dcb40f331fff1ab10257
hacking is not capped in g-r and it is in
blacklist for requirement as hacking new version
can break the gate jobs.
Hacking can break gate jobs because of various
reasons:
- There might be new rule addition in hacking
- Some rules becomes default from non-default
- Updates in pycodestyle etc
That was the main reason it was not added in g-r
auto sync also. Most of the project maintained the
compatible and cap the hacking version in
test-requirements.txt and update to new version when
project is ready. Bumping new version might need code
fix also on project side depends on what new in that
version.
If project does not have cap the hacking version then,
there is possibility of gate failure whenever new hacking
version is released by QA team.
Example of such failure in recent release of hacking 1.1.0
- http://lists.openstack.org/pipermail/openstack-dev/2018-May/130282.html
Change-Id: I05790c5f946a4127fe6c8d9867cc618cb83d67fb
This patch enforces the removal of the container
network namespace when the environment is destroyed.
Closes-Bug: #1769651
Change-Id: I9b0bbbb5a59e6067a745635c555051ef97b79f9a
Signed-off-by: Jorge Niedbalski <jorge.niedbalski@linaro.org>
Allow to set computes_need_external_bridge to true/false
depending on the desired configuration, for allowing
cases such as disable dvr and enable l3 ha.
Closes-Bug: #1769686
Change-Id: I1565b08dfccb7bec2ddda8c048b7d951c9eb1824
Signed-off-by: Jorge Niedbalski <jorge.niedbalski@linaro.org>
Some variables should be modified too, just like this commit:
https://review.openstack.org/552863
Change-Id: Id43eefb948a05de03d19c46048b6151bf6f5dc0a
Related-Bug: #1769463
Aodh service credential defaults to publicURL.
Aodh should work with the internal interface (v3 API Identity syntax):
.....
[service_credentials]
interface = internal
.....
Change-Id: Ie11290be25291fee1938032ae6294546fb7a98f3
In ansible/roles/etcd/tasks/config.yml, the kolla_docker
compare_container action doesn't check environment.
Once a container is created, it won't get recreated if only the
environment change. This commit add the environment attribute to the
kolla_docker action in etcd role
Change-Id: I8fb71cc945867e06acc67f6d1256bf62f4276206
Closes-Bug: #1765517
Magnum was unable to fire up k8s cluster because heat-container-agent
inside kube-master was pointing to internal keystone endpoint instead of
public endpoint. This fix tells kolla ansible to set clients_keystone
auth_uri to public endpoint so that heat-container-agent communication
with heat is successfully authenticated by keystone.
Change-Id: Ida49528f88685710b5e6b8f3c4d4622506af5ae1
Closes-Bug: #1762754
Since the kolla-kubernetes has been separated, and nowhere to
use kubetoolbox, it is safe to remove it from kolla-ansible.
Change-Id: I9aa59d45fa04c13944120f1f99af2361551147ad
Keystone removed uuid token provider in Rocky
This patch change the default value and fix comments for the option.
Change-Id: Idca0004852b688fcdd34ef47c38dec6b8bf05f86
Closes-Bug: #1757520
Assuming both Prometheus and Grafana are enabled, this change
configures Grafana such that it can use data collected by the
Prometheus server.
Partially-Implements: blueprint prometheus
Change-Id: I0290806d0c622b1de33cf7047f099795b4c780b9
The documentation for integration with external Ceph refers to
enable_manila_backend_ceph_native where it should be
enable_manila_backend_cephfs_native.
Change-Id: If727166e61a8e911445630dc6d6adaa2126fa7e8
The baremetal scheduling options were deprecated
in Pike and the ironic_host_manager was deprecated
in Queens and has now been removed. Deployments
must use resource classes now for baremetal scheduling.
Depends-On: https://review.openstack.org/565805/
Change-Id: I92cab1c7b5dc8201b0a8a94287c75d6348ab4465
Setting endpoint_type to internalURL in the murano config allows murano
to work properly when using self-signed certificates.
Change-Id: I5873f92994d790d597ec5e97163bbf8001c362b9
