Patch - code.engineering.redhat.com/gerrit/#/c/200373
Testproject Patch with run - https://code.engineering.redhat.com/gerrit/#/c/199982/
Patch - code.engineering.redhat.com/gerrit/#/c/200505
Testproject with job run - code.engineering.redhat.com/gerrit/#/c/200636
Patch - https://code.engineering.redhat.com/gerrit/#/c/200511/
Testproject with job run - https://code.engineering.redhat.com/gerrit/#/c/200637/
Patch - code.engineering.redhat.com/gerrit/#/c/200673
Testproject with job run - code.engineering.redhat.com/gerrit/#/c/200674
Fixed after moving from validate-tempest to os_tempest
Patch - https://code.engineering.redhat.com/gerrit/#/c/201430/
Testproject with job run - https://code.engineering.redhat.com/gerrit/#/c/200637/
Track of issue faced earlier - rabbitmq can't start properly in some sceanrios when ipv6 is entirely disabled on the system. ipv6 was getting disabled due to downstream nodepool specific customization.
After fix - https://code.engineering.redhat.com/gerrit/#/c/201431/1 and the new image builds - https://sf.hosted.upshift.rdu2.redhat.com/nodepool-log/cloud-rhel-8-1-0000013385.log - we are good
Patch - https://code.engineering.redhat.com/gerrit/#/c/201376/
testproject - https://code.engineering.redhat.com/gerrit/#/c/201382/
Failing with emit_releases_file.py: error: argument –stable-release: invalid choice: 'rhos-17
2020-05-26 06:33:35.825941 | primary | usage: emit_releases_file.py [-h] --stable-release
2020-05-26 06:33:35.826019 | primary | {newton,ocata,pike,queens,rocky,stein,train,ussuri,master}
2020-05-26 06:33:35.826028 | primary | --distro-name {centos} --distro-version {7,8}
2020-05-26 06:33:35.826035 | primary | --featureset-file FEATURESET_FILE
2020-05-26 06:33:35.826040 | primary | [--output-file OUTPUT_FILE] [--log-file LOG_FILE]
2020-05-26 06:33:35.826045 | primary | [--upgrade-from] [--is-periodic]
2020-05-26 06:33:35.826061 | primary | emit_releases_file.py: error: argument --stable-release: invalid choice: 'rhos-17' (choose from 'newton', 'ocata', 'pike', 'queens', 'rocky', 'stein', 'train', 'ussuri', 'master')
It looks like the internal jobs were not looking for the internal ceph containers in the correct locations.
The review above try to address that.
NOTE: tripleo-ci-internal tenant went down on Monady 05/18.
This was CIX'ed. If you catch infra, you can rerun and check.
Note also the reparenting from the job that would be in the integration line (to match the upstream RDO pattern).
@rlandy Thank you for reparenting.
* Marios also pointed considering using periodic-tripleo-ci-rhel-8-standalone-component-rhos-17 as parent to get osp-17 release vars
* https://code.engineering.redhat.com/gerrit/#/c/200373/2/zuul.d/component-jobs.yaml
* https://code.engineering.redhat.com/gerrit/#/c/200373/5/zuul.d/component-jobs.yaml
and same thought hit me as well but at that time i thought the approach I used earlier was slight better - Thinking about future releases when rhos18/19 will comes, we can just reuse "periodic-tripleo-ci-rhel-8-scenario001-standalone-component "(if we keep it generic ) as a parent job to create
- periodic-tripleo-ci-rhel-8-scenario001-standalone-component-rhos-17 ( passing release/branch related vars there)
- periodic-tripleo-ci-rhel-8-scenario001-standalone-component-rhos-18
- periodic-tripleo-ci-rhel-8-scenario001-standalone-component-rhos-19
But I am okay with your approach if this keep things consistent with RDO.
My scenario1 Patch need slight changes.. I made them and rechecked your testpatch.
@rlandy @marios fyi.. Gathered some thoughts here on possible parenting design.. Will need your suggestions to help in opting the best possible solution here. (We are still currently focusing on sceanrio001 as rest other sceanrios will follow the pattern - scenario001 currently is running good but failing due to prometheues containers)
Context:-
In rdo/upstream whenever we create a new scenario job - we create it for master and whenever there is branching(new release) - we inherit from master for the new branched job(for ex: ussuri).
example:-
But in downstream there is no Master to follow , Right now in downstream(In our initial work[1]) we are treating rhos-17 as master. If we follow same pattern when RHOSP18 will come - in order to not repeat ourselves we can possibily use rhos-17 as parent and override release specific vars(This can just work fine) - but this can create a challenge if in future we decided to drop rhos17 jobs.
[1] https://code.engineering.redhat.com/gerrit/#/c/200373/14/zuul.d/standalone-jobs.yaml
What can be done -
First possible solution:-
I am wondering if in downstream we should first just create generic jobs(without any release specific vars) instead of treating 17 as master, This way we can just reuse the generic jobs to create rhos17/18/19.. jobs(as they come). We will diverge from RDO/Upstream pattern but it seems needed as we don't have a master to follow.
I brainstormed to rework https://code.engineering.redhat.com/gerrit/#/c/200373/14/zuul.d/standalone-jobs.yaml to something like below:-
standalone-jobs.yaml
175 # openstack-periodic-rhos-17 jobs
176 - job:
177 name: periodic-tripleo-ci-rhel-8-standalone-rhos-17
178 parent: tripleo-ci-rhel-8-standalone-internal-base-dlrn
179 # This need to change to branch used for integration
180 branches: .*
181 voting: false
182 pre-run:
183 - playbooks/configure-rhel.yaml
184 - playbooks/pre-rhos.yaml
185 vars: &vars_rhos17 ----> Add an anchor here
186 dlrnapi_host: "osp-trunk.hosted.upshift.rdu2.redhat.com"
187 dlrnapi_protocol: "http"
188 release: master
189 osp_release: "rhos-17"
190 branch_override: "stable/rhos-17"
191 rhos_release_version: 17
192 release_file: promotion-testing-hash-rhos-17
193 promote_source: tripleo-ci-testing
194 featureset_override:
195 tempest_tempest_conf_overrides:
196 image.image_path: 'http://images.rdoproject.org/cirros/cirros-0.4.0-x86_64-disk.img'
Create a generic new scenario001 job without any rhosp-17 vars.
- job:
name: periodic-tripleo-ci-rhel-8-scenario001-standalone
+- parent: tripleo-ci-rhel-8-standalone-internal-base-dlrn --> ( changed from periodic-tripleo-ci-rhel-8-standalone-rhos-17)
+ branches: .*
+ voting: false
+ pre-run:
+ - playbooks/configure-rhel.yaml
+ - playbooks/pre-rhos.yaml
+ abstract: true
vars:
standalone_ceph: true
featureset_override:
standalone_environment_files:
- 'environments/low-memory-usage.yaml'
- 'ci/environments/scenario001-standalone.yaml'
tempest_services:
- aodh
tempest_run_concurrency: 1
tempest_tempest_conf_overrides:
'telemetry.alarm_granularity': '60'
tempest_test_whitelist:
- 'tempest.api.identity.v3'
- 'tempest.scenario.test_volume_boot_pattern.TestVolumeBootPattern.test_volume_boot_pattern'
- 'tempest.scenario.test_snapshot_pattern.TestSnapshotPattern.test_snapshot_pattern'
- 'telemetry_tempest_plugin.scenario.test_telemetry_integration.TestTelemetryIntegration'
- job:
name: periodic-tripleo-ci-rhel-8-scenario001-standalone-rhos-17
parent: periodic-tripleo-ci-rhel-8-scenario001-standalone
vars: *vars_rhos17
Changes in component-jobs.yaml
# scenario001-standalone-component
- job:
name: periodic-tripleo-ci-rhel-8-scenario001-standalone-component
parent: periodic-tripleo-ci-rhel-8-scenario001-standalone ------> (changed from periodic-tripleo-ci-rhel-8-scenario001-standalone-rhos-17)
abstract: true
vars: *component_vars
- job:
name: periodic-tripleo-ci-rhel-8-scenario001-standalone-component-rhos-17
parent: periodic-tripleo-ci-rhel-8-scenario001-standalone-component
abstract: true
vars: *component_vars_rhos17
- job:
name: periodic-tripleo-ci-rhel-8-scenario001-standalone-cloudops-rhos-17
parent: periodic-tripleo-ci-rhel-8-scenario001-standalone-component-rhos-17
vars:
component: cloudops
Alternative Possible Solution
https://code.engineering.redhat.com/gerrit/#/c/200373/14/zuul.d/standalone-jobs.yaml
job:
name: periodic-tripleo-ci-rhel-8-scenario001-standalone-rhos-17
parent: periodic-tripleo-ci-rhel-8-standalone-rhos-17
vars: &scen1_component_vars ----> Add an anchor here
standalone_ceph: true
featureset_override:
standalone_environment_files:
- 'environments/low-memory-usage.yaml'
- 'ci/environments/scenario001-standalone.yaml'
tempest_services:
- aodh
tempest_run_concurrency: 1
tempest_tempest_conf_overrides:
'telemetry.alarm_granularity': '60'
tempest_test_whitelist:
- 'tempest.api.identity.v3'
- 'tempest.scenario.test_volume_boot_pattern.TestVolumeBootPattern.test_volume_boot_pattern'
- 'tempest.scenario.test_snapshot_pattern.TestSnapshotPattern.test_snapshot_pattern'
- 'telemetry_tempest_plugin.scenario.test_telemetry_integration.TestTelemetryIntegration'
job:
name: periodic-tripleo-ci-rhel-8-scenario001-standalone-rhos-18
parent: periodic-tripleo-ci-rhel-8-standalone-rhos-18
vars: *scen1_component_vars
This can create a challenge if in future we decided to drop rhos17 jobs.
What can be done if above situation comes in future?