# Component pipeline Baremetal Scenarios
###### tags: `Design`

## How baremetal jobs in component pipeline can be a game changer?

* Triplei-ci Job's - In my opinion we are the second layer of defence(First in my opinion are ci jobs defined in each projects- nova, neutron) to defend against issue and bugs entering the Openstack world.
Exception - For tripleo we are also the first layer of defence.

* We as the second layer are crucial because no other component ci before - is actually testing the real deployment which is tripleo currently(traditionally  - the rpm based + now services are containerized but its tripleo who is still preparing those containers for other components)

* We are also testing the integration between different component.

But still bugs sometimes go into product and few causes, Why:-

1) We are not always testing all the scenarios(not feasible)
2) We are not testing scenarios in undercloud/overcloud but in standalone

Proof:- https://bugs.launchpad.net/tripleo/+bug/1881420

Ex: My GSS colleague recently hit an issue with Octavia and reached out to me to know if we are seeing similiar issue in CI.

On standlone those tasks which failed on undercloud was skipped:-

https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_341/725250/9/gate/tripleo-ci-centos-8-scenario010-standalone/3418fb5/logs/undercloud/home/zuul/standalone_deploy.log

~~~
bc764e20-0ff8-80b0-1cbd-00000000016d | SKIPPED | Make needed directories on the undercloud | undercloud
bc764e20-0ff8-80b0-1cbd-00000000016e | TASK | Write group_vars file
bc764e20-0ff8-80b0-1cbd-00000000016d | TIMING | Make needed directories on the undercloud | 0:11:49.657 | 0.10s
bc764e20-0ff8-80b0-1cbd-00000000016e | SKIPPED | Write group_vars file | undercloud
~~~

**Solution** :- I understand we cannot test all the sceanario in tripleo(undercloud/overcloud) based environment because of resource constraint. But may be we should test them somewhere/ anywhere? I know one possible place :)

**16.2 Component pipeline is up, What's next?**
* We are still not testing real customer production scenario but now with baremetal jobs in 16.2 pipeline we can do it to some extent?

**Do we need one/more big fat customer like scenario - in downstream?** to capture bugs which we are missing in standalone job?

What can we possibly include in one big fat job?

- Minions
- Composable roles
- Multi nic
- TLS
- Octavia
- DVR
- DPDK / SRIOV (If hardware permits)
- Service telemetry framework (If Openshift bits available)
- Ceph Nodes
- HCI
- Scale down
- Cinder multiple backend (lvm, NFS, Ceph)
- Update
- Upgrade
- Scale up
- Testing Live migrations of instances

**Brainstorming about some Complex Customer like Scenarios**
https://docs.google.com/spreadsheets/d/1UC9BeLdoxNv1SWolOapr4TyEhCC5TFcuff_2eUUVUG0/edit#gid=0