K8s multi. Admin guide. New OpenShift deployment docs #391

ghost · 2018-04-13T06:53:58Z

Do not merge until eclipse-che/che#9190 is merged

What's done:

Kubernetes admin guide. Common for OpenShift and K8s.
OpenShift config and K8s config pages refer to admin guide
Deploy with help to K8s docs
Updated instructions on how to deploy to OpenShift - no scripts

slemeur

Really nice content.

Should we have a page "OpenShift - Admin Guide" just linking to the kubernetes one?

slemeur · 2018-04-13T08:17:47Z

src/main/pages/setup-kubernetes/kubernetes-admin-guide.md

+
+You can get the list of available versions at [Che GitHub page](https://github.com/eclipse/che/tags).
+
+Since `nightly` is the default tag used in Che deployment, and image pull policy is set to Always, triggering a new deployment, will pull a newer image, if available.


probably add link on how to change the image pull policy

Done.

As to link to K8S admin guide, it is references from OpenShift Config page.

bmicklea

I'm not sold on the idea of forcing OpenShift users to jump through the K8S and OS install pages. GitHub pages lets you use common text to populate multiple pages - we should use something like that so the K8S and OpenShift pages can have identical content sections (updated in one place) without forcing one or other to read multiple pages.

bmicklea · 2018-04-16T21:12:51Z

src/main/pages/overview/infra-support.md

-| installers               | yes          | some installers may require [sudo access](openshift-config.html#enable-ssh-and-sudo)       |
-| file system permissions  | not limited  | limited to directories owned by root [group](openshift-config.html#filesystem-permissions) |
+|  Feature                 | **Docker**   | **OpenShift**                                                                              | **Kubernetes** |
+| root access              | yes          | no (See: [Configuration](openshift-config.html#enable-ssh-and-sudo))                       | yes |


I think this line should say "yes" for OpenShift with the link you provided since you can do it.

bmicklea · 2018-04-16T21:13:21Z

src/main/pages/overview/infra-support.md

+| root access              | yes          | no (See: [Configuration](openshift-config.html#enable-ssh-and-sudo))                       | yes |
+| https                    | no           | yes (See: [Configuration](openshift-config.html#https-mode))                               | yes |
+| scalability              | no           | yes (See: [Configuration](openshift-config.html#scalability))                              | yes |
+| priviliged containers    | yes          | no (configurable in [OpenShift](https://docs.openshift.com/container-platform/3.6/admin_guide/manage_scc.html#grant-access-to-the-privileged-scc))         | yes |


same as above, since you can do it - should say "yes"

bmicklea · 2018-04-16T21:13:58Z

src/main/pages/overview/infra-support.md

+| priviliged containers    | yes          | no (configurable in [OpenShift](https://docs.openshift.com/container-platform/3.6/admin_guide/manage_scc.html#grant-access-to-the-privileged-scc))         | yes |
+| health checks            | no           | yes                                                                                        | yes |
+| persistent preview URLs  | no           | yes                                                                                        | yes |
+| installers               | yes          | some installers may require [sudo access](openshift-config.html#enable-ssh-and-sudo)       | yes |


Again - change text for OpenShift to "yes, some installers may require..."

bmicklea · 2018-04-16T21:15:55Z

src/main/pages/setup-kubernetes/kubernetes-admin-guide.md

+
+**Single User**
+
+Che server pod consumes up to 1GB RAM. The initial request is 256MB, and server pod rarely consumes more than 800MB. A typical workspace will require 2GB. So, **<span style="color:red;">3GB</span>** is a bare minimum to try single user Che on OpenShift/Kubernetes.


remove "bare" (not needed)

bmicklea · 2018-04-16T21:19:49Z

src/main/pages/setup-kubernetes/kubernetes-admin-guide.md

+
+* If there's `CHE_INFRA_KUBERNETES_OAUTH__TOKEN` fabric8 client will use it. If both are available, token has a priority over username/pass.
+
+* If none are available **che service account** is used to create workspace objects. This is default behavior. Since che SA cannot create objects outside a namespace it's bound to, all workspace objects are created in Che namespace. Admins can grant `che` [service account super user permissions](https://kubernetes.io/docs/admin/authorization/rbac/#service-account-permissions), and this way it will be possible to use this SA to create objects outside Che namespace. OpenShift cluster admins can do it this way using oc CLI:


I don't understand what "Since che SA cannot..." means - is that a typo?

SA == service account. Fixed.

bmicklea · 2018-04-16T21:20:57Z

src/main/pages/setup-kubernetes/kubernetes-admin-guide.md

+|**common**|One PVC for all workspaces, sub-paths pre-created| easy to manage and control storage. no need to recycle PVs when pod with pvc is deleted | ws pods should all be in one namespace
+|**unique**|PVC per workspace| Storage isolation | An  undefined number of PVs is required |
+
+## Common PVC strategy


Headings should consistently use title case.

bmicklea · 2018-04-16T21:23:21Z

src/main/pages/setup-openshift/openshift-config.md


-This is going to be changed once [this issue](https://github.com/eclipse/che/issues/8178) is completed.
+This page describes OpenShift specific configuration. Refer to [Kubernetes Admin Guide][kubernetes-admin-guide] to general information that works both for OS and K8S.


So for OpenShift someone would do everything in the K8S section, then come back here and do all the stuff here?

Fixed. Now k8s guide is embedded in k8s and openshift admin guide pages

ghost · 2018-04-17T17:30:32Z

@bmicklea all issues addressed. PR in Che merged. Merging this one as well.

Eugene Ivantsov added 5 commits April 10, 2018 17:38

Admin guide

6ecf497

Admin guide

1c493f1

k8s

2ce6477

Admin guide. New deployment stuff

7794441

Cleanup

77b1b58

ghost requested review from bmicklea and slemeur as code owners April 13, 2018 06:53

benoitf added the status/code-review label Apr 13, 2018

slemeur approved these changes Apr 13, 2018

View reviewed changes

Eugene Ivantsov added 2 commits April 13, 2018 11:53

pull policy

5f5337a

Kubernetes in supported platforms

eae6934

bmicklea reviewed Apr 16, 2018

View reviewed changes

Eugene Ivantsov added 2 commits April 17, 2018 08:56

Cleanup. Embedd k8s guide into k8s and openshift pages.

5d79ec7

Cleanup. Embedd k8s guide into k8s and openshift pages.

a22e44e

ghost mentioned this pull request Apr 17, 2018

Use templates only to deploy Che to OpenShift eclipse-che/che#9190

Merged

k8s fix

c96cbe2

ghost merged commit f1c1d3e into master Apr 17, 2018

ghost deleted the admin_guide branch April 17, 2018 17:30

benoitf removed the status/code-review label Apr 17, 2018

This pull request was closed.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

K8s multi. Admin guide. New OpenShift deployment docs #391

K8s multi. Admin guide. New OpenShift deployment docs #391

ghost commented Apr 13, 2018

slemeur left a comment

slemeur Apr 13, 2018

ghost Apr 13, 2018

ghost Apr 13, 2018

bmicklea left a comment

bmicklea Apr 16, 2018

ghost Apr 17, 2018

bmicklea Apr 16, 2018

ghost Apr 17, 2018

bmicklea Apr 16, 2018

ghost Apr 17, 2018

bmicklea Apr 16, 2018

ghost Apr 17, 2018

bmicklea Apr 16, 2018

ghost Apr 17, 2018

bmicklea Apr 16, 2018

ghost Apr 17, 2018

bmicklea Apr 16, 2018

ghost Apr 17, 2018

ghost commented Apr 17, 2018


		You can get the list of available versions at [Che GitHub page](https://github.com/eclipse/che/tags).

		Since `nightly` is the default tag used in Che deployment, and image pull policy is set to Always, triggering a new deployment, will pull a newer image, if available.


		Single User

		Che server pod consumes up to 1GB RAM. The initial request is 256MB, and server pod rarely consumes more than 800MB. A typical workspace will require 2GB. So, <span style="color:red;">3GB</span> is a bare minimum to try single user Che on OpenShift/Kubernetes.


		* If there's `CHE_INFRA_KUBERNETES_OAUTH__TOKEN` fabric8 client will use it. If both are available, token has a priority over username/pass.

		* If none are available che service account is used to create workspace objects. This is default behavior. Since che SA cannot create objects outside a namespace it's bound to, all workspace objects are created in Che namespace. Admins can grant `che` [service account super user permissions](https://kubernetes.io/docs/admin/authorization/rbac/#service-account-permissions), and this way it will be possible to use this SA to create objects outside Che namespace. OpenShift cluster admins can do it this way using oc CLI:


		This is going to be changed once [this issue](https://github.com/eclipse/che/issues/8178) is completed.
		This page describes OpenShift specific configuration. Refer to [Kubernetes Admin Guide][kubernetes-admin-guide] to general information that works both for OS and K8S.

K8s multi. Admin guide. New OpenShift deployment docs #391

K8s multi. Admin guide. New OpenShift deployment docs #391

Conversation

ghost commented Apr 13, 2018

slemeur left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bmicklea left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ghost commented Apr 17, 2018