Jtesar/chapter1 web console install #1
Conversation
|
modules/chapter1/pages/section1.adoc
Outdated
| *Red{nbsp}Hat Openshift Data Science* is installed as an operator. As a dependecy it requires the *Red{nbsp}Hat Openshift Pipelines* operator to be installed as well. | ||
|
|
||
| == Installation of the Red{nbsp}Hat Openshift Pipelines operator |
There was a problem hiding this comment.
Please see how Jaime is structuring the sections in the first intro course. Add 2 or 3 paragraphs about the context for the exercise.You can talk about the following things mentioned in the design doc
- [Trevor] - This seems to focus on the self-managed install but doesn't talk at all about the managed service. The difference between the two should be discussed somewhere in the training. Introduce the different ways you can install the product - CLI, web console etc. Talk about how different self managed is compared to managed RHODS. Separate product docs exist for the AWS managed version of the product. You can say this is out of scope for this course and that we will only cover self managed.
- Provide warnings about not choosing the community upstream OpenDataHub operators
- [Trevor] RHODS includes other dependencies outside of OpenShift Pipelines. For GPU we require the NVIDIA GPU Operator, and the Node Feature Discovery Operator.
modules/chapter1/pages/section1.adoc
Outdated
|
|
||
| == Installation of the Red{nbsp}Hat Openshift Pipelines operator | ||
|
|
There was a problem hiding this comment.
See and talk to Jaime about splitting this section into a separate sub-sections called "Exercise: Installing RHODS using the web console" and another sub-section called "Installing RHODS using the CLI". A third sub-section called "Troubleshooting" can be added. Finally, add a References section and provide links to all product docs etc plus any code and scripts you have written for CLI install. Provide links to appropriate scripts/code repo
There was a problem hiding this comment.
Ignore my comment about separate sub section for CLI install. I notice that you have added a separate section which is the better approach because of the length of the exercise. But please add a Troubleshooting and References section
modules/chapter1/pages/section1.adoc
Outdated
| When you click on the *Red{nbsp}Hat Openshift Data Science* dashboard button a login window should appear. | ||
| + | ||
| image::rhods_verify2.png[width=800] |
There was a problem hiding this comment.
Mention the fact that they should log in as cluster administrator using OCP credentials to do this initial checking of install status
modules/chapter1/pages/section1.adoc
Outdated
| = Installation using the Web Console | ||
|
|
||
| *Red{nbsp}Hat Openshift Data Science* is installed as an operator. As a dependecy it requires the *Red{nbsp}Hat Openshift Pipelines* operator to be installed as well. |
There was a problem hiding this comment.
You might also mention (maybe as a note) that if they are planning to use GPUs, they must have the NVIDIA GPU operator installed too.
modules/chapter1/pages/section1.adoc
Outdated
|
|
||
| 1. Login to Red{nbsp}Hat Openshift using a user which has the cluster-admin role assigned. | ||
| 2. Navigate to **Operators** -> **OperatorHub** and seach for *Red{nbsp}Hat Openshift Pipelines* |
There was a problem hiding this comment.
(adoc) There is a menu macro that you can use for this kind of navigation
| 2. Navigate to **Operators** -> **OperatorHub** and seach for *Red{nbsp}Hat Openshift Pipelines* | |
| 2. Navigate to menu:Operators[OperatorHub] and search for *Red{nbsp}Hat Openshift Pipelines*. |
https://docs.asciidoctor.org/asciidoc/latest/macros/ui-macros/#menu-macro-syntax
There was a problem hiding this comment.
To use the menu macro, you must activate the experimental attribute in the project, like I did here https://github.com/RedHatQuickCourses/rhods-intro/blob/main/antora-playbook.yml#L11
modules/chapter1/pages/section1.adoc
Outdated
|
|
||
|
|
||
| 4. In the installation view some installation parameters can be tuned. Administrator can set the *Update{nbsp}channel* parameter to a specific version and the *Update{nbsp}approval* parameter to either *Automatic* or *Manual*. The *Installation{nbsp}mode* and the *Installed{nbsp}namespace* parameters are fixed. |
There was a problem hiding this comment.
For UI buttons or other elements, you can use the button macro https://docs.asciidoctor.org/asciidoc/latest/macros/ui-macros/#button-macro-syntax
modules/chapter1/pages/section1.adoc
Outdated
| image::pipeline_install2.png[width=800] | ||
| + | ||
| Click on the **Install** button at the bottom of to view the to proceed with the installation. A window showing the installation progress will pop up. |
There was a problem hiding this comment.
Example of use of the button macro (you have to enable experimental mode for this to work).
| Click on the **Install** button at the bottom of to view the to proceed with the installation. A window showing the installation progress will pop up. | |
| Click on the btn:[Install] button at the bottom to proceed with the installation. A window showing the installation progress will pop up. |
modules/chapter1/pages/section1.adoc
Outdated
| image::pipeline_install3.png[width=800] | ||
| + | ||
| When the operator is installed you can view it's details by clicking on **View{nbsp}Operator** button. |
There was a problem hiding this comment.
| When the operator is installed you can view it's details by clicking on **View{nbsp}Operator** button. | |
| When the operator is installed you can view its details by clicking on **View{nbsp}Operator** button. |
modules/chapter1/pages/section2.adoc
Outdated
| = Section 2 | ||
| = Installation using the Web Console | ||
|
|
||
| *Red{nbsp}Hat Openshift Data Science* is installed as an operator. As a dependecy it requires the *Red{nbsp}Hat Openshift Pipelines* operator to be installed as well. |
There was a problem hiding this comment.
OpenShift Pipelines is only required if you plan to install the Data Science Pipelines component.
There was a problem hiding this comment.
True, but the v1 operator installs the Data Science Pipelines component automatically so I see the pipelines operator as a dependency for v1. What do you think?
modules/chapter1/pages/section2.adoc
Outdated
| image::rhods_install2.png[width=800] | ||
|
|
||
| 4. In the installation view window choose the **Update Channel**, **Installed{nbsp}Namespace** and *Update approval** or accept default values and click on **Install* the button. The *Installation{nbsp}mode* parameter is fixed. |
There was a problem hiding this comment.
The default channel today installs the v1 operator which will be depreciated in the near future. The last version of v1 operator should be released in the next week or two and v2 will be the default moving forward.
V2 is available in the alpha channel today but we be made available in stable before the end of the year.
There was a problem hiding this comment.
Yes, I know that. I asked specifically about whether we create the module for v1 or v2, because installation for v2 would look quite differently. The workshop is in November and v1 will still be the stable version then I guess . So I think we should finish the initial version of the modeule using v1 and update it when v2 becomes stable.
There was a problem hiding this comment.
Yeah, the current state of the transition of the operator does leave us in a bit of an awkward position.
Personally I still think the docs should be oriented towards the v2 operator. The goal of the workshop is to enable internal resources to help deliver RHODS for customers. By the time the people attending the workshop are actively working on an engagement with RHODS, v2 will be stable.
It doesn't seem like the best strategy to try to enable people with material and knowledge that will be out of data a month later.
At a minimum, for the workshop we should at least be discussing v2 and letting people know that it is coming. If the workshop is going to focus on v1 for the hands on stuff that is fine.
I would probably defer to Noel on what he thinks on this topic of how to handle this for the workshop.
Just as a general update around v2, it is very usable today. The biggest piece that is missing that is holding it back from stable is the automation intended to handle the v1 to v2 migration.
There was a problem hiding this comment.
Yeah, that's a fair point, I agree. I'll somehow include the v2 as well :-)
| image::rhods_install3.png[width=800] | ||
| + | ||
| Operator Installation progress window will pup up. The installation may take a couple of minutes. |
There was a problem hiding this comment.
With the v1 operator as part of the install the operator pod automatically installs all of the components in the background as part of the operator install which is why it takes so long.
With v2, only the operator is installed and a new object called a DataScienceCluster is a new required object that must be setup. The "installing operator" screen will be a little bit different because of this required object and the user will have to take additional steps to create the DSC after the operator is installed.
The DSC will allow users to enable and disable specific components and have better control over what is installed on each cluster.
There was a problem hiding this comment.
I'd reply the same thing as to the previous comment. We should finalize the first version using v1 and then update to v2 once it becomes stable.
Initial version of Web console installation