Improve ceph update run tasks by detecting if ceph update is required
to not start ceph update to the same version. Also exclude ceph images
from container check when update of those containers is not happening.
Change-Id: Idd4132c75208b635c9bf06e28457e02aeb314708
The directory is created under root so we need to become one. The
overcloud command is fine as /etc/openstack/cloud.yaml is correctly
defined in CI.
Change-Id: I289ecaa7ca7bdf7b295593dd430f513e8c24b3f8
Currently, the jobs only collect log at the end of the update.
Add some log collection during the update to the /var/log/extra
directory of each host to get information before the update and during
the update to ease debugging.
Change-Id: I2c4b59b2ab8f8b0d64c9982b528d927d3f6f2436
This is first attempt to test whole upgrade procedure from
Train to Wallaby. We for now deal just with the OpenStack
upgrade part without OS upgrade step. When we will understand
the process better we will refactor this code.
Changes:
- Remove tls templates comversion.
- Cleanup adjust-roles-data. We might introduce new ones.
- Handle the network switch to ansible templates
- For now use Fast Furious procedure to generate playbook
- Add ceph adoption workflow from ceph-ansible to cephadm
Co-authored-by: jbadiapa@redhat.com
Change-Id: Iae1c9ef1244676589ee505201093daa52704a438
Add file to the reno documentation build to show release notes for
stable/zed.
Use pbr instruction to increment the minor version number
automatically so that master versions are higher than the versions on
stable/zed.
Sem-Ver: feature
Change-Id: I625ba2be680c5f2091966e8285b041cdc015bf35
When a group is a child of another group then the child groups role mappings
are added to the parent group role mappings. This which fails with a
KeyError when the child is an empty group as no mapping is created.
Closes-bug: #1994081
Related: rhbz#2089512
Change-Id: I05f4989b0f14d56d3c33f91b3cccb85b729987cb
Previously volume creation was fast enough, but it doesn't seem to be
anymore. Add also a loop sleep during the deletion.
Change-Id: I88c73a7789c2246a12b80e52829f8f2f1a0a10e5
`merge-new-params-nic-config-script.py`from Tripleo Heat Templates
is no longer supported and will be removed for Zed release
along with the `convert_nic_config.py`.
Tasks using these scripts have to be be removed as well.
This includes indirect usage, and inclusion of the tasks
in FFU process.
Related: https://review.opendev.org/c/openstack/tripleo-heat-templates/+/860691
Signed-off-by: Jiri Podivin <jpodivin@redhat.com>
Change-Id: If769954dea48cba5b9cc52871a18735e8520a0ab
Add more error detection in workload_launch script to fail script early
when the error occurs.
Also, make ssh test to run by default to be sure created vm is
accessible via floating ip, before saving its floating ip to a file. The
file is used in ping test and we need to make sure it is not present on
undercloud when vm fails to boot or is unreachable.
Change-Id: I1057de8dde62fef7e2e00880704d7f3bb3bf6b19
If overcloud is deployed with fencing enabled, then
we should temporarily disable fencing for the duration of the update.
We are introducing a pre-undercloud update step to disable
it and post-update step to enable it back.
Change-Id: Ie6324fc2c9cdbeac6126e5bdbbce41f23e143be5
To be able to run selectable validations ansible tasks, additional tags
are added to all update related validations tasks.
By introducing this change one can run only specific validations with
ansible tags mechanism.
These new tags are introduced:
- pre_update_validations
- pre_update_prepare_validations
- pre_update_run_validations
- post_update_validations
Change-Id: I57d6680f16b2bdc9026a3f8175470f35611a6b45
* Remove repomap.csv as it was removed by leapp (f09c8add)
* Add an actual list of files
* Add retires and delay to improve resilience over networking issue
Change-Id: I710f2703b57de330cacf6d1e631180f4c0f13fba
Append item to custom_env_files if it's already present in
undercloud.conf
Insert custom_env_files when undercloud.conf doesn't have it
Closes-Bug: 1981081
Change-Id: I0d62ad4e10e1dbb2a6019815fc133237f6ee40fe
Instead of 'cephadm shell -- ceph -s -f json | jq .progress_events' we should
use 'cephadm shell -- ceph orch upgrade status' to monitor progress of ceph
update to determine (1) whether an upgrade is in progress.
[1] https://docs.ceph.com/en/quincy/cephadm/upgrade/#monitoring-the-upgrade
Change-Id: If19485304508b5d2979ddc7c6050ea87b06a99c6
From wallaby+ Ceph is deployed by cephadm and minor updates
are considered day2 operations that can be performed using the
cephadm cli [1].
This patch adjusts current automation from the tripleo-upgrade
project for ceph update related tasks.
`overcloud external-update run --tags ceph` command is replaced with
cephadm. The 'ceph' tags were removed from tht[2] and ceph update can't be
run with TripleO commands. Cephadm should be used instead to safely upgrade
Ceph from one release to the next.
[1] https://docs.ceph.com/en/pacific/cephadm/upgrade/#starting-the-upgrade
[2] https://review.opendev.org/c/openstack/tripleo-heat-templates/+/830051
Change-Id: I5fe5c10b9f997b12cb87a24cbf00a48f2f6cca8c
Ie02259336c763cfa980b59ad64f9e478d05690fb dropped
the script which is used everywhere.
Closes-Bug: #1979707
Change-Id: I55e29ba86a847a14a363c224055885bfd086fbb6
The tripleo-ansible-inventory needs the authentication information in
the environment and the inventory option doesn't work on train.
Change-Id: Iae27f89e64a9276017a453db194e9f357ef338e2
Closes-Bug: 1973743
Tripleo-ci jobs are broken after latest release of setuptools 61.0
because of breaking changes which are not backwork compatible,
details in related bug and [1].
Users that don't set ``packages``, ``py_modules``, or ``configuration`` are
still likely to observe the auto-discovery behavior, which may halt the
build if the project contains multiple directories and/or multiple Python
files directly under the project root.
To disable auto discovery, one can do below in setup.py
~~~
setuptools.setup(..,packages=[],..)
~~~
or
~~~
setuptools.setup(..,py_modules=[],..)
~~~
Also, adding ^setup.*$ under irrelevant-files as content-provider
parent[2] also have same.
[1] https://github.com/pypa/setuptools/issues/3197
[2] https://opendev.org/openstack/tripleo-ci/src/branch/master/zuul.d/base-upstream.yaml#L92
Releated-Bug: #1966382
Change-Id: I643952b64d557f3bd81699e505588133cfe02f52
It's possible to also use --networks-file instead of
-n in the overcloud deploy command. This change adds
a secondary search string to the awk command to ensure
we capture the file regardless of which argument has
been used in the overcloud deploy script.
Resolves: rhbz#2064354
Change-Id: I2387d508f9dc91dc71c19ae3625fe093d33c2bc0
Make sure compute role nodes are updated after the entire control
plane has been updated even in composable role and cephstorage role
nodes are updated last.
In composable deployment the control plane is sliced up in controller,
database, messaging. Currently we only ensure that controller get
updated first, so messaging or networker for instance get updated after
compute.
This kind of disruption in the control plane can lead to unexpected
issue on the compute role. Furthermore they can be seen as false
positive as the compute role would be updated after the all the
Controller roles.
Change-Id: Ib3bf87510763edc86f2300e3b28708f6ef38ea18