Since we are working in process of removing tempest container
from wallaby release and all the tempest run is switched to
os_tempest for all supported release in tripleo-ci.
This change is not going to affect any periodic jobs also.
It also removes run_tempest and tempest_format var.
Related-Bug: #1916875
Signed-off-by: Chandan Kumar (raukadah) <chkumar@redhat.com>
Change-Id: I3d046047833044fd2d75ba399f784be5d992fbef
This configuration should not be used in featureset, because it would
never be overwriten, it should be in the playbooks/tempest.yml in
tripleo-quickstart-extras
Change-Id: I06480c30e0c671c3f97361c8bdc06c5fc28dfb94
The current default value is literraly 'omit' which is incorrect.
If a users doesn't have any intention to override any tempestconf
values, the overrides should be an empty dict or not set at all [1].
The review sets the default value of tempestconf overrides to
an empty dict as shown in the doc [1].
[1] https://docs.openstack.org/python-tempestconf/latest/user/profile.html
Change-Id: Ifd791bae7982154fd097774a3992530675278345
Remove *_container_cli settings out of
feature set file to single-source
the setting in tqe role extras-common.
Change-Id: I833348a18597c230fea4a01ab203569ffe5ef7a0
Depends-On: https://review.opendev.org/#/c/678881/
pacemaker and other critical settings have
diverged from 20/21. This aligns the two.
Closes-Bug: #1823347
Change-Id: I5da36a06a34969abde6ea63cdaf75118d4eee4b4
Although featureset021 is supposed to run full tempest without skip
list, some plugins doesn't have support into tripleo or
python-tempestconf. In this case, these plugins tests are always get
executed making the job run slower, and in sometimes the controller just
hang out due the huge amount of unecessary requests. This also make it
hard to get a real baseline because featureset021 will always fail
because of these unconfigured/unsupported plugins.
Change-Id: I800a2850a71017b90ad89e17051563691b509dbf
This adds ctlplane_masquerade: true for f21 which is used by
periodic-tripleo-ci-centos-7-ovb-1ctlr_1comp-featureset021-master
As reported in the bug below this job fails as the compute node
can't reach pool.ntp.org
Change-Id: Ib146eaaefb202f1303454b5b86a33b6de8a096f1
Closes-Bug: 1794258
fs021 is switched to containerized undercloud in [1],
only fs021 is the ovb job that has missing this net
config, this patch fixes it.
[1] https://review.openstack.org/#/c/583202
Related-Bug: #1749707
Change-Id: I773041057bd5bc19fc280ef7ccb445bbccfa30a3
Ensure the latest CentOS base package updates and
minor release updates do not break tripleo
Closes-Bug: #1788675
Change-Id: Ida0b644c9216eebe1d38ee72ead3da569d941f39
Tempest container is available and built when we deploy
containerized undercloud.
Since undercloud is containerized from rocky and ongoing, this patch
enable containerized tempest when undercloud is running as container.
Also, moved all tempest related configuration to the end of file to
avoid overwritten options
Depends-On: https://review.openstack.org/#/c/580384/
Change-Id: Icdb0f2b033ad0d1ba2548c591fa32dc649fff319
We have ntp_args which include ntp_server variable. Don't set it
in extra agrs so we can overwrite it.
Change-Id: Ib6a30b4581f095b89f0a128232c165c53a716a48
Bug was introduced in https://review.openstack.org/#/c/566527/
Set default value for tempest_format in logs collection role
when it publishes docs.
Closes-Bug: #1776164
Change-Id: I25a73bbd3804cf013f38627ca3d742e3a6a2f5d1
As we move towards making config-download the default, this converts the
remaining ovb jobs in the check queue that run against master changes to
use config-download.
Once it's the default, we can come back and update these configs to not
have to explicitly use --config-download.
This also updates the feature configuration doc to indicate these ovb
jobs and the multinode jobs from the previous patch now use
config-download.
Change-Id: Ib0c0ba67e13f0978017f67dfc6095221d79b244c
implements: blueprint config-download-default
As we backported ci network configs to all stable branches
including newton, we don't need to use tripleo-ci templates.
Change-Id: Id94d2e41c31b66f90e878336e72caff59b33d1ca
Update quickstart ovb-ha (ipv4) jobs to use the network templates and
environments from tripleo-heat-templates instead of tripleo-ci. This
allows for having per-branch differences between the templates.
The net-iso.yaml environment from tripleo-ci is uneeded as it's
parameters are also set in network-environment.yaml, so it's usage is
removed.
Since we are not backporting all the nic config templates for ci to the
newton branch of tripleo-heat-templates, the network_isolation_args
variable is special cased to still use the templates from tripleo-ci for
the newton branch.
Change-Id: Iabad6df081555300e2c793b323acf42ea17dc5d4
Move all network settings to environment OVB file in CI,
leaving only features here.
Depends-On: I695b3eddb821821bfba0566dfa599e5a5b264d6a
Change-Id: I7345318bcba633f5384313c25be0716d0a425070
Fix network isolation args as it should be according to changes
in https://review.openstack.org/#/c/419049/ that fixes
failures of ovb-ha oooq job.
Change-Id: I059207581b6e867458433aeeea118106815fed71
This patch adds blank newline for each [yaml|sh] files if they
don't end with one. This will allows us getting linters happy and avoid us
having git adding "\ No newline at end of file" in diffs when pushing
a new patch.
In other way, you can configure Git to properly handle line endings in
typing:
$ git config --global core.autocrlf input
Change-Id: I8a113541b3eae6a36ed32995822dedbbd2d9666e
Signed-off-by: Gael Chamoulaud <gchamoul@redhat.com>
This change takes the overcloud_nodes variable and defines that
under config/nodes/foo.yml. This depend on the change
to quickstart-extras where a definition of topology (nodes)
will be created. Pass the topology in an explicit variable vs.
using extra_args.
The original configuration files in config/general_config/ will
continue to work but are now deprecated.
The featureset files that define standard set of feature to test are
also added, and should correspond to test configuration for jobs that
are typically found upstream
Change-Id: Id6dd3f71b33f76a0401fe8a84080fe80de0dbb69
Co-Authored-By: Gabriele Cerami <gcerami@redhat.com>
We introduced a few network isolation types support in
patch https://review.openstack.org/#/c/437437.
Now it's required to set type of network isolation.
Set centos.ci network isolation type by default to single nic
with vlans which was used always by quickstart before.
Change-Id: If3ad22a2db0cb9420264745620aa57841da5f1dd
This patch is replacing the usage of the pxe_ssh driver in Ironic in
favor of pxe_ipmitool + VirtualBMC*. The main reason for this patch is
because the pxe_ssh driver is already marked as unsupported and will be
removed from Ironic upstream at the end of the Ocata release.
Older releases: Liberty, Mitaka and Newton will continue to rely on the
pxe_ssh driver.
* VirtualBMC is a proxy that converts IPMI commands to libvirt calls
Partial-Bug: #1645733
Change-Id: Iaaa9fc1fd593cbce045456f4341c461a8eea82ad
Review https://review.openstack.org/#/c/397312/ has removed
some options in tripleoclient.
Specifically, --neutron-network-type and --neutron-tunnel-types
are being used in tripleo-quickstart. This patch removes them
from extra_args.
This parameters are being passed via environment files or using
vxlan by default so i think it doesn't need any other adjustment.
Change-Id: I8f9b52e7a0e78266fab0717e20a9710be264f5aa
By default, undercloud is created with 8 vcpus, while compute and
controller are created only with one. This make the controller
struggle with requests from tempest that runs on undercloud, and
have the concurrency set to vcpus/2, which means 4 threads doing
a lot of requests to controller and this one have only one vcpu
to handle these requests, plus all the services running on the vm.
In my tests, 2 vcpus were enough to make a full tempest run pass
without bottlenecks.
Also keeping the same configuration for HA, since in this case
the workload is spread, so there's no need to change the actual
setup for now
Change-Id: Idbdff31e77b239a4c5ec5e7bf918b89a0a6cf533
Adding artcl_create_docs_payload to the ha toplogy and enabling
post-build doc creation by tripleo-collect-logs
Change-Id: Idda252f08e1d4de15929af22f45ae52ae82478b1
Three out of the four chassis' in CentOS CI have slower CPUs, but
they have more of cores. I have tested bumping the number of CPUs
on the undercloud on these slower chassis' and it made a big
difference. More so, the memory pressure on the undercloud is
actually not bad now, so there is not too much risk in hitting
issues where more CPUs causes high memory consumption.
This also increases the default in the HA configuration by the same
factor (2x).
Change-Id: I44f34827774353f9f70e7fd51c524567914059ab
'minimal' deployments have begun to fail with the existing, 60 minute,
default time. Bumping to prevent build failures
- Increase default timeout to 75 minutes
- Remove timeout override for HA
Change-Id: I03f0592a528313ae22caf130e8234173db4ef4ba
Run pingtest in minimal configuration instead of tempest
Change parameter name in HA configuration to common ones
https://bugzilla.redhat.com/show_bug.cgi?id=1344339
Change-Id: I8945a6c0e571021ee13ca09481781a9299d13e9f
* move the roles, libraries and config to a top level dir
* adjust the ansible.cfg to account for the new role path
* create a config directory
config/release <- only for release and image_url info
config/general_config <- everything else
* update collect-logs role path
* add ci-scripts/full-minimal.sh so that jjb will not have
to be updated after this.
Change-Id: I2e6129553bca253fa47ce56bfcbc2f244e74409e
When deploying with `network_isolation: true`, we are failing to pass
in the network settings from our configuration into the tripleo
deployment process, leading ultimately to deployment failures when
attempting to use non-default values.
additionally, setting `network_isolation: true` required adding a
number of explicit arguments to `extra_args`, so it was not possible
to control the use of network isolation simply by toggling the value
of `network_isolation`.
This commits allows the deployment to succeed with `network_isolation`
either disabled or enabled, and allows the use of non-default values
for the network configuration.
Change-Id: I68f5a87e390aa2ea8d8c5631eba17dc5dd44270e
Closes-bug: #1581659
this commit modifies the `extra_args` definitions in a couple of
variable files to use YAML blocks, rather than a really long quoted
string. This makes the files display better in a number of situations
(nobody likes massic horizontal scrolls), and it makes the arguments a
little more obvious.
Change-Id: I98ed41013ad4a72f43b74ff95fd04a42a845d636
This adds an ovs port on the undercloud on the external network.
Adds the required network-environment.yaml file for net-iso template
deployment. Update the centos/min.yml and centos/ha.yml w/ network
isolation templates and created a new settings file to deploy w/o
network isolation.
Change-Id: If1841710c0e9d9f59a81410e91f460a8088f020c
Marios Andreou
Chandan Kumar (raukadah)
Arx Cruz
Martin Kopec
Ronelle Landy
Arx Cruz
Wes Hayutin
Sorin Sbarnea
yatin
Bogdan Dobrelya
Sagi Shnaidman
James Slagle
Emilien Macchi
Gael Chamoulaud
Lucas Alvares Gomes
Alfredo Moralejo
Juan Antonio Osorio Robles
Harry Rybacki
John Trowbridge
Yolanda Robla Mota
Mathieu Bultel
Lars Kellogg-Stedman