Split Dynamic Resource Allocation into a Resource Management section - #56295
Conversation
👷 Deploy Preview for kubernetes-io-vnext-staging processing.
|
✅ Pull request preview available for checkingBuilt without sensitive environment variables
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
/label refactor |
…allocation/_index.md Co-authored-by: Tim Bannister <193443691+lmktfy@users.noreply.github.com>
|
Given ++ from the wg-device-management WG https://kubernetes.slack.com/archives/C0409NGC1TK/p1782721132858409 This is a good first step and ready to merge. next will be to repoint broken links |
johnbelamaric
left a comment
There was a problem hiding this comment.
some tweaks - though I am also OK merging this as-is and then applying these changes in a follow up PR
| api_metadata: | ||
| - apiVersion: "resource.k8s.io/v1alpha3" | ||
| kind: "DeviceTaintRule" | ||
| - apiVersion: "resource.k8s.io/v1beta2" |
| is not exempt either. An admin using that mode must explicitly tolerate all taints | ||
| to access tainted devices. | ||
|
|
||
| Device taints and tolerations is controlled by the |
There was a problem hiding this comment.
We can remove this paragraph as it is going GA
| Consider the following example: | ||
|
|
||
| ```yaml | ||
| apiVersion: scheduling.k8s.io/v1alpha2 |
There was a problem hiding this comment.
This is moving to beta, we probably should update the version here
There was a problem hiding this comment.
Some part of the site still uses scheduling.k8s.io/v1alpha2 (podgroup-api and gang-scheduling docs), so bumping just this example would be inconsistent, is there a tracking change for the PodGroup beta graduation I should follow?
|
|
||
| <!-- body --> | ||
|
|
||
| ## Extended resource allocation by DRA {#extended-resource} |
There was a problem hiding this comment.
this feature doesn't really require any special driver support, maybe it should move up to dra-api.md, or we should remove the comment that these features require driver support
| The resulting ResourceClaim will contain a request for an `ExactCount` of the | ||
| specified number of devices of that DeviceClass. | ||
|
|
||
| Extended resource allocation by DRA is controlled by the |
| If a non-multiply-allocatable device were chosen, the allocation would have resulted in the entire device. | ||
| To force the use of a only multiply-allocatable devices, you can use the CEL criteria `device.allowMultipleAllocations == true`. | ||
|
|
||
| ### DistinctAttribute constraint |
There was a problem hiding this comment.
this also does not need driver support
There was a problem hiding this comment.
I will update the overview from them requiring driver
| in the `kube-apiserver` and `kube-controller-manager`. | ||
|
|
||
|
|
||
| ## DRA device metadata in containers {#device-metadata} |
There was a problem hiding this comment.
this belongs in dra-features.md not here
…allocation/_index.md Co-authored-by: John Belamaric <jbelamaric@google.com>
…allocation/_index.md Co-authored-by: John Belamaric <jbelamaric@google.com>
…nagement-section # Conflicts: # content/en/docs/concepts/scheduling-eviction/dynamic-resource-allocation.md
… into dra-resource-management-section
…allocation/device-taints.md Co-authored-by: John Belamaric <jbelamaric@google.com>
…allocation/device-taints.md Co-authored-by: John Belamaric <jbelamaric@google.com>
|
This is not a change to the live site, so I can Lazy consensus because it is a big reorganization. Anyone is welcome to unhold this after 48 hours, or an approver for the English localizations can unhold at any time. |
|
LGTM label has been added. DetailsGit tree hash: fffd5ff0396d960c6ffa88e41092f80a2f90c4e6 |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: lmktfy The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
Thanks @johnbelamaric I will make these fixes as a follow-up rather than here, to keep the Lazy consensus. Once this merges, I suggest #56265 (device taints GA) goes in first, then the follow-up on top. |
|
@pohly , what do you think of the reorg of the DRA docs? |
|
Lazy consensus expired with no actual objection. /unhold |
Adds a
Concepts > Resource Managementsection and splits the Dynamic Resource Allocation page into purpose-built pages:Approach: verbatim content move, then dropped the
alpha features/beta featuresheadings, each capability now sits on its topical page with its{{< feature-state >}}badge carrying the stage, so graduating features get ahome instead of a maturity bucket to reshuffle each release.
Redirects:
aliaseson the DRA landing preserve the old/concepts/scheduling-eviction/dynamic-resource-allocation/URL.Scope: structure only. Inbound-link repointing and the v1.37 graduation content follow in separate PRs.
Issue
Related to #49899
discussion on slack here https://kubernetes.slack.com/archives/C1J0BPD2M/p1782232560810309