Notes: demonstrate python dynamic-client usage & brainstorming strategic-merge-patch, deployment rolling restart method, tekton pipelines, Container Storage Interface (CSI) session #4
May 3, 2021
Notes from today:
- Improved this PR (add examples to demonstrate the usage of dynamic client #1448) from yesterday with review pointers. What I fixed is the following:
apiextensions.k8s.io/v1in place of
apiextensions.k8s.io/v1betafor CustomResourceDefintion (CRD) api
- changing the
v1made adding versions schema mandatory. So, learnt a lot today about writing CRD schema.
- improved variables names
ingressroute_manifest_first, so it become more consistent with
ingressroute_manifest_second(well yes, I’m silly to name this as
*-second, in place of one of two, or first or second)
- as suggested, tried using
application/strategic-merge-patch+jsonwhile applying patches to
custom resource (CR)objects, but there is some issue here. Seems like, the
strategic-merge-patchis not supported. Atleast that’s what the error says precisely. But this patch is for supporting the
strategic merge patch. See this comment here for context. I’ll wait for more input on this from the reviewers.
The reviewer to one of the previous PR (add example to demonstrate a rolling restart of the deployment #1450) suggested to replicate the same example for a k8s python
typed client. So, I refactored & cleaned a previous
deployment_crud.pypython script to include a
rolling_restart_deploymentmethod. Here is the PR (update deployment_crud.py to include a
rolling_restart_deploymentmethod for typed client #1452)
- During the office hours, I worked on writing some openshift tekton pipelines/tasks (which are a part of our team’s central autonomous platform,
url_shortnertekton task, which enhances the google-chat notification messages from
github/ci-pipeline(another tekton pipeline) by adding a shortened url of the running
pipelinerunlogs tab. PR here.
report-unresolved-pagerduty-incidentstekton pipeline, which will send a report of
unresolvedPagerDuty incidents on the respective (team’s) google-chat channel, after a schedule of every 6 hours (which is at the end of every IC on-call shifts). WIP PR here (I’m still testing it).
- spent time setting up a development test namespace to perform testing on new tasks/pipelines defined under
- Lastly, joined a k8s community session around Container Storage Interface (CSI). Well, I hardly understood anything there, but that’s my inexperience with the concept. The session was very thorough. I’m keeping the notes & recording here for myself.
That’s all from today.