List operations should add the task_type criterion to the user
This patch also fixes the neutron-tempest-plugin-designate-scenario
job to use the ussuri version neutron-tempest-plugin-designate-scenario-ussuri.
This is required for this patch to merge.
(cherry picked from commit fb9f0b33d3)
This change enforces the update of the zone masters
for all zones that belongs to a particular pool, using
the pool's defined target(s) masters and forcing a update_zone
This change also, moves the backend base class update_zone
method as an abstract method, allowing to each backend
implementation to create its own update logic. For the
case of bind9 its extended to allow running a rndc modzone
with the new given masters for the zone fixing the behavior
exposed on LP: #1879798.
Signed-off-by: Jorge Niedbalski <email@example.com>
(cherry picked from commit 3756fc51e7)
We no longer need to deal with xenial or older distro versions, so we
can assume that we can always install pdns4 from the distro directly
and don't need the distro version check any longer.
Drop a config option that was removed in pdns 4.2.x.
Use mysql credentials correctly.
(cherry picked from commit 6f17fe10bb)
This should prevent us from obsessing over zones that are in the
designate DB but not in the pdns backend due to races/incomplete
(cherry picked from commit 5acba9443e)
When running the lower-constraints test we are currently
using pip install -e . to install the source. This caused
python libraries to unintentionally getting upgraded beyond
the intended constraints.
(cherry picked from commit 568b559b67)
we create a default pool when exec database sync, and get the
default_pool_id from designate.conf, but we did not register this
opt before using it, so we always get the default value which is
(cherry picked from commit 609437e73d)
If no notification handler is configured the
designate-sink may not shutdown properly.
This is because the listener was created but never
started. This caused oslo.messaging to hang on
waiting for a successful start before stopping.
(cherry picked from commit 40cbc9bd43)
Update the URL to the upper-constraints file to point to the redirect
rule on releases.openstack.org so that anyone working on this branch
will switch to the correct upper-constraints list automatically when
the requirements repository branches.
Until the requirements repository has as stable/ussuri branch, tests will
continue to use the upper-constraints list on master.
This patch standardizes the CONTRIBUTING.rst file and adds information
from the required doc/source/contributor/contributing.rst.
Signed-off-by: Nicolas Bock <firstname.lastname@example.org>
The current locking implementation is limited to
the running process. This introduces distributed
locking that will help prevent race conditions
when there are many instances of designate-central
Make a few cleanups:
- Remove python 2.7 stanza from setup.py
- Add requires on python >= 3.6 to setup.cfg so that pypi and pip
know about the requirement
- Remove obsolete sections from setup.cfg:
* Wheel is not needed for python 3 only repo
* Some other sections are obsolete
- Update classifiers
- Update requirements, no need for python_version anymore
The repo is Python 3 now, so update hacking to version 3.0 which
supports Python 3.
Fix problems found.
Update local hacking checks for new flake8.
# to unbreak gate:
There is a potential race condition when running
the unit tests with how eventlet is loaded. This
patch changes the neutron api implementation to use
GreenThreadPoolExecutor to make sure that
it does not fail depending on when eventlet applies
it's monkey patching.
The use of this code path is very limited at the moment,
and this patch further simplifies the code to make it easier
to either deprecate or extend it in the future.
* Renamed service_status.py -> heartbeat_emitter.py.
* Merged HeartbeatEmitter and Heartbeat.
* Mark Akamai v2 as untested
* Mark Akamai eDNS as known broken due to the API being removed
* Add basic docs for Akamai v2
Signed-off-by: Graham Hayes <email@example.com>
Work on fixing the postgres job has stalled for more than half a year
now, so stop running this known-broken job on every patch. This can be
reverted once the job is fixed.
Tempest recently moved tox default
to py3 including all-plugin env.
This will fail the py2 jobs which are using
all-plugin tox env. Because tox all-plugin
env var enable the sitepackages=True. Py2 jobs
install the tempest plugins on py2 and all-plugin
tox try to find the installed plugins on py3 which
cause no test to run and fail the job.
tox all-plugin has been deprecated since long and not
supposed to use for plugins who already in migrated to
separate repo in queens cycle.
Using tox 'all' env to fix the issue.
Legacy jobs except grenade job has been migrated
to zuulv3 in - https://review.opendev.org/#/c/644586
But legacy playbooks has been left to remove. Those are
unused now and can be removed from master and train branch.
This is a bit of a security issue, also we don't seem to need it anymore
since we have long switched to running under python3 and the setcap
doesn't have any effect anyway. It will also break deployments once we
drop installing python2 completely, so just remove it now.
The USE_SYSTEMD variable is to be dropped from devstack, as running
with systemd is the only option now. Drop it's usage in our plugin.
The new heartbeat code was causing the service to
not exit properly. This patch removes the use of a
thread group for handling heartbeats and replaces
it with a FixedIntervalLoopingCall instead.
- Ignore duplicate Zone error
- Handle error when contractId or gid is missed
- Ignore port for masters servers, because Akamai uses only 53 port and
does not allow to specify any port in list of masters servers.
- Added timeout and retries for soft Zone Delete
- Added handling errors on the delete zone action
- Added Log info message with RequestId on soft zone delete
- Added processing for TsigKey during creation zone
- Added devsatck_plugin for akamai_v2 backend
The network_api endpoint code would never
actually use the manually configured endpoints.
This patch changes that by first checking if
custom endpoints were configured, and if that
isn't the case it will use the service_catalog
In addition we introduce full testing on
this code path to make sure it behaves
as we expect.
Reduced the number of calls we need to make when creating records using
the sink by better using the create/update recordset api.
This also fixes a bug where the sink could trigger a race condition in
the worker causing it to throw a BadAction exception.