Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Reducing the installation options for Che #20178

Closed
2 tasks done
l0rd opened this issue Jul 22, 2021 · 13 comments
Closed
2 tasks done

Reducing the installation options for Che #20178

l0rd opened this issue Jul 22, 2021 · 13 comments
Labels
area/che-operator Issues and PRs related to Eclipse Che Kubernetes Operator area/chectl Issues related to chectl, the CLI of Che area/install Issues related to installation, including offline/air gap and initial setup kind/enhancement A feature request - must adhere to the feature request template. kind/epic A long-lived, PM-driven feature request. Must include a checklist of items that must be completed. roadmap/3-months Epics that are planned to complete in the short term (within 3 months) severity/P1 Has a major impact to usage or development of the system.

Comments

@l0rd
Copy link
Contributor

l0rd commented Jul 22, 2021

Is your enhancement related to a problem? Please describe.

We currently have multiple options to deploy Che. That's hard to maintain and it doesn't bring value for customers.

Describe the solution you'd like

Platforms Current Options New Options (step1) New Options (step2)
minikube/k8s/microk8s/docker-desktop helm/olm/operator operator operator
openshift 4/crc helm/olm/operator olm none (not supported)
minishift operator operator operator

A few clarifications:

  • The goal is to have one installation mode only. That's easier to maintain and simpler for users. The only installation mode would be operator (applying the Kubernetes resources to deploy Che operator CRD and controller without OLM).
  • On OpenShift there is the option to install CRW (Red Hat distribution of Che) and that should be the default choice. That's why we want to remove Che support for olm (step 2).
  • We want to keeep olm support temporarly (step 1) untlil we have weekly releases of CRW (downstream issue). That's because we want have a rapid validating of OLM installation (nightlies build rather than 6 weeks build cadence of CRW).
  • we replace current helm chart with a new one that uses operator templates. We publish the new chart to https://artifacthub.io/ and we setup a gh action to automatically release a new version of the chart at every che release

Installers for k8s platform:

k8sinstallers

Tasks:

Subepic:

@l0rd l0rd added the kind/enhancement A feature request - must adhere to the feature request template. label Jul 22, 2021
@che-bot che-bot added the status/need-triage An issue that needs to be prioritized by the curator responsible for the triage. See https://github. label Jul 22, 2021
@l0rd l0rd added kind/epic A long-lived, PM-driven feature request. Must include a checklist of items that must be completed. roadmap/6-months Epics that are planned to complete in the medium term (within 6 months) and removed status/need-triage An issue that needs to be prioritized by the curator responsible for the triage. See https://github. labels Jul 22, 2021
@l0rd
Copy link
Contributor Author

l0rd commented Jul 22, 2021

@tolusha please create a subtask for step 1 and add it as sprint/next. That includes stop supporting helm charts so it's a big change. We need to deprecate it initially and, if there are no strong push back, disable it after a couple of sprints.

@tolusha
Copy link
Contributor

tolusha commented Jul 22, 2021

Helmchart deprecation
#20183

@l0rd l0rd added area/install Issues related to installation, including offline/air gap and initial setup area/che-operator Issues and PRs related to Eclipse Che Kubernetes Operator area/chectl Issues related to chectl, the CLI of Che area/ci CI build and releases, PR testing, & whitelabel/productization issues roadmap/3-months Epics that are planned to complete in the short term (within 3 months) and removed roadmap/6-months Epics that are planned to complete in the medium term (within 6 months) labels Sep 29, 2021
@rhopp
Copy link
Contributor

rhopp commented Oct 12, 2021

@l0rd Small nitpick - In your table, shouldn't be minishift moved to the row with minikube/k8s/microk8s/docker-desktop? AFAIK minishift doesn't have OLM capabilities (to satisfy step 1).

@l0rd
Copy link
Contributor Author

l0rd commented Oct 12, 2021

@rhopp thanks. Fixed

@nickboldt nickboldt added the severity/P1 Has a major impact to usage or development of the system. label Oct 14, 2021
@nickboldt
Copy link
Contributor

Setting P1 since this is an item to be done in the next 3 mo.

@gazarenkov
Copy link
Contributor

What is the role of chectl in this context?
I see it mentioned as the area/chectl label but no mentions in the description.
Does the operator option automatically imply using chectl (only?) as a mechanism to perform it?

@l0rd
Copy link
Contributor Author

l0rd commented Oct 15, 2021

chectl sever:deploy is the way to deploy Che using the command line. And it should continue to be like that.

Step 1 implies that we should deprecate the "--installer" parameter because there is only one per platform. Step 2 (blocked by this downstream issue) means that we should deprecate "openshift" and "crc" platforms as those should be better served/covered by CRW. @tolusha does that make sense for you?

@nickboldt
Copy link
Contributor

nickboldt commented Nov 5, 2021

On OpenShift there is the option to install CRW (Red Hat distribution of Che) and that should be the default choice. That's why we want to remove Che support for olm (step 2).

Does this mean that crwctl will ONLY be used for OCP 3.11 (operator mode) ?

Or is the plan to keep olm mode in crwctl, but remove it from chectl?

Asking because QE relies heavily on crwctl to do install tests w/ CRW 2.13. Seems a strange choice to make crwctl an Officially Supported install method, then remove it a few months later.

So if crwctl will continue support olm mode, we need to be clear about HOW we'll maintain the fork downstream once we get to step 2.

@tolusha
Copy link
Contributor

tolusha commented Nov 8, 2021

OCP 3.11: crwctl with --installer operator
OCP 4.x: OpeatorHub + crwctl with --installer olm

@nickboldt
Copy link
Contributor

Ah, so no change then. This is the current state of affairs for CRW:

OCP 3.11: crwctl with --installer operator
OCP 4.x: OpeatorHub + crwctl with --installer olm

@nickboldt nickboldt removed the area/ci CI build and releases, PR testing, & whitelabel/productization issues label Nov 23, 2021
@Kasturi1820
Copy link

@tolusha Is step 1 done ?

@tolusha
Copy link
Contributor

tolusha commented Jan 6, 2022

@Kasturi1820
Yes, completed.
Besides the fact, that it is still possible to use --installer operator to deploy on OpenShift (moslty for testing purpose)

@l0rd
Copy link
Contributor Author

l0rd commented Jan 6, 2022

Closing this issue as the first step is done (reducing the installation options). The second step (remove Che from OperatorHub) is tracked in a distinct issue and is blocked by this downstream issue.

@l0rd l0rd closed this as completed Jan 6, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
area/che-operator Issues and PRs related to Eclipse Che Kubernetes Operator area/chectl Issues related to chectl, the CLI of Che area/install Issues related to installation, including offline/air gap and initial setup kind/enhancement A feature request - must adhere to the feature request template. kind/epic A long-lived, PM-driven feature request. Must include a checklist of items that must be completed. roadmap/3-months Epics that are planned to complete in the short term (within 3 months) severity/P1 Has a major impact to usage or development of the system.
Projects
None yet
Development

No branches or pull requests

7 participants