Uploaded image for project: 'Hybrid Application Console'
  1. Hybrid Application Console
  2. HAC-2979

[Hypershift ROSA][Details] Call node pool endpoint when adding a machine pool for Hypershift clusters

XMLWordPrintable

    • 3
    • False
    • False
    • None
    • HAC Infra OCM - Sprint 231, HAC Infra OCM - Sprint 233, HAC Infra OCM - Sprint 234

      User story:

      So that I can manage my cluster, as a user I want the ability to add additional machine (aka node) pools to an existing cluster.

      Acceptance criteria:

      *. A user is able to create a node pool for a hypershift control plane cluster.

      • A user must select a valid node pool name (same validation as ROSA cluster machine pools)
      • A user must select a private subnet that is attached to their account when adding a machine pool to a hypershift cluster.
      • For non-hypershift clusters, there isn't an option to add a subnet when adding a machine pool (current functionality)

      Mockups:

      PD 1464

      -------------------------------------------------------------------------------------------------------------------------------

      Implementation details

      Summary:

      The main task is that if the cluster is a hypershift cluser, the node pool POST endpoint is called instead of the[ machine pool endpoint|https://api.openshift.com/#/default/post_api_clusters_mgmt_v1_clusters_cluster_id_machine_pools].

      More information needed

      • There are numerous differences between the machine pool endpoint and the node pool endpoint.

      *On Machine pool end point, but not on the node pool endpoint*

      • `aws` endpoint (see below)
      • `availabilty_zones` => this is a single availablity_zone (array vs single) in hypershift
      • `instance _type`
      • `labels` (see below)
      • `security_group_filters`
      • `subnets` => this is a single subnet (array vs single) in hypershift
      • `taints`

      *On Node pools, but missign on machine pool endpoint*

      • `auto_repair`

      *`aws` vs `aws_node_pool`*
      This appears to be the main information about the need pools sent to the back end. There are numerous differences between the two.

      The main question is what information gathered in the current UI can be re-used for a node-pool and what items need to be added/hidden for hypershift.

        1. image-2023-02-09-12-17-24-882.png
          207 kB
          Thi Le
        2. Screenshot 2023-01-19 at 4.06.33 PM.png
          114 kB
          Kim Doberstein
        3. Screenshot 2023-01-19 at 4.06.57 PM.png
          225 kB
          Kim Doberstein

            kdoberst Kim Doberstein
            kdoberst Kim Doberstein
            Jayakrishnan Mekkattillam Jayakrishnan Mekkattillam
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: