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

Document extended resources and OIR deprecation. #5399

Merged
merged 2 commits into from
Sep 13, 2017
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
110 changes: 110 additions & 0 deletions docs/concepts/configuration/manage-compute-resources-container.md
Original file line number Diff line number Diff line change
Expand Up @@ -305,6 +305,15 @@ where `OOM` stands for Out Of Memory.

## Opaque integer resources (Alpha feature)

{% include feature-state-deprecated.md %}

**Note:** Opaque Integer Resources will be removed in version 1.9.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[Extended Resources](#extended-resources) are a replacement for Opaque Integer
Resources. Users can use any domain name prefix outside of the `kubernetes.io/`
domain instead of the previous `pod.alpha.kubernetes.io/opaque-int-resource-`
prefix.
{: .note}

Kubernetes version 1.5 introduces Opaque integer resources. Opaque
integer resources allow cluster operators to advertise new node-level
resources that would be otherwise unknown to the system.
Expand All @@ -316,6 +325,7 @@ available amount is simultaneously allocated to Pods.
**Note:** Opaque integer resources are Alpha in Kubernetes version 1.5.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we can drop this note entirely now.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

agreed

Only resource accounting is implemented; node-level isolation is still
under active development.
{: .note}

Opaque integer resources are resources that begin with the prefix
`pod.alpha.kubernetes.io/opaque-int-resource-`. The API server
Expand Down Expand Up @@ -395,6 +405,106 @@ spec:
pod.alpha.kubernetes.io/opaque-int-resource-foo: 1
```

## Extended Resources

Kubernetes version 1.8 introduces Extended Resources. Extended Resources are
fully-qualified resource names outside the `kubernetes.io` domain. Extended
Resources allow cluster operators to advertise new node-level resources that
would be otherwise unknown to the system. Extended Resource quantities must be
integers and cannot be overcommitted.

Users can consume Extended Resources in Pod specs just like CPU and memory.
The scheduler takes care of the resource accounting so that no more than the
available amount is simultaneously allocated to Pods.

The API server restricts quantities of Extended Resources to whole numbers.
Examples of _valid_ quantities are `3`, `3000m` and `3Ki`. Examples of
_invalid_ quantities are `0.5` and `1500m`.

**Note:** Extended Resources replace [Opaque Integer
Resources](#opaque-integer-resources-alpha-feature). Users can use any domain
name prefix outside of the `kubernetes.io/` domain instead of the previous
`pod.alpha.kubernetes.io/opaque-int-resource-` prefix.
{: .note}

There are two steps required to use Extended Resources. First, the
cluster operator must advertise a per-node opaque resource on one or more

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

s/opaque/extended or s/opaque//

nodes. Second, users must request the Extended Resource in Pods.

To advertise a new Extended Resource, the cluster operator should
submit a `PATCH` HTTP request to the API server to specify the available
quantity in the `status.capacity` for a node in the cluster. After this
operation, the node's `status.capacity` will include a new resource. The
`status.allocatable` field is updated automatically with the new resource
asynchronously by the kubelet. Note that because the scheduler uses the
node `status.allocatable` value when evaluating Pod fitness, there may
be a short delay between patching the node capacity with a new resource and the
first pod that requests the resource to be scheduled on that node.

**Example:**

Here is an HTTP request that advertises five "example.com/foo" resources on
node `k8s-node-1` whose master is `k8s-master`.

```http
PATCH /api/v1/nodes/k8s-node-1/status HTTP/1.1
Accept: application/json
Content-Type: application/json-patch+json
Host: k8s-master:8080

[
{
"op": "add",
"path": "/status/capacity/example.com~1foo",
"value": "5"
}
]
```
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This http request detail is not strictly necessary because the following Shell command already covers all necessary information. So ... I'd suggest we remove it.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK, should we remove the similar one from the OIR section above too?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If there is no meat in it, better throw it away IMO.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I suggest to remain it. It's very important for usual users.
I can help user to config this kind of resource though shell.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@zhangxiaoyu-zidif The snippet above is not a JSON or an YAML for you to create resource from Shell. It is an explanation of the HTTP request internals, and the information is duplicated with the command below.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry to misunderstand your response. The command below is necessary. And the above one can be remove. It has no meaning to common users. But the below shell command is useful. =)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

My historical note is that I first added the HTTP request, then reviewers asked for a curl equivalent. Guess the rationale is that the K8s API speaks HTTP and not curl. But I'm apparently in the minority here. Will remove both of them.


```shell
curl --header "Content-Type: application/json-patch+json" \
--request PATCH \
--data '[{"op": "add", "path": "/status/capacity/example.com~1foo", "value": "5"}]' \
http://k8s-master:8080/api/v1/nodes/k8s-node-1/status
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the '/status' part should be removed?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ConnorDoyle
I think at least 50% common user do not know how to config OIR or Extended Resouce by curl quickly. So remaining it is very necessary. I meet many users ask how to config OIR, so I add it in OIR docs. =)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@tengqm the URL shown is the correct way to patch the node status.

```

**Note**: In the preceding request, `~1` is the encoding for the character `/`
in the patch path. The operation path value in JSON-Patch is interpreted as a
JSON-Pointer. For more details, see
[IETF RFC 6901, section 3](https://tools.ietf.org/html/rfc6901#section-3).
{: .note}

To consume an Extended Resource in a Pod, include the resource name as a key
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

better be explicit here: .. include the resource name example.com/foo as a key ...

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I meant this sentence to be more general than the specific resource from the example. Did that not come across? This is the second of two steps required to use ERs. The previous section began with "To advertise a new Extended Resource..."

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Alright, if it is referring to general cases, i.e. "an Extended Resource", it is fine.
I was reading it as "the Extended Resource".

in the `spec.containers[].resources.requests` map.

**Note:** Extended resources cannot be overcommitted, so request and limit
must be equal if both are present in a container spec.
{: .note}

The Pod is scheduled only if all of the resource requests are
satisfied, including cpu, memory and any Extended Resources. The Pod will
remain in the `PENDING` state as long as the resource request cannot be met by
any node.

**Example:**

The Pod below requests 2 cpus and 1 "example.com/foo" (an extended resource.)

```yaml
apiVersion: v1
kind: Pod
metadata:
name: my-pod
spec:
containers:
- name: my-container
image: myimage
resources:
requests:
cpu: 2
example.com/foo: 1
```

## Planned Improvements

Kubernetes version 1.5 only allows resource quantities to be specified on a
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ This page shows how to specify opaque integer resources for a Node.
Opaque integer resources allow cluster administrators to advertise node-level
resources that would otherwise be unknown to Kubernetes.

{% include feature-state-alpha.md %}
{% include feature-state-deprecated.md %}

{% endcapture %}

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ title: Assign Opaque Integer Resources to a Container

This page shows how to assign opaque integer resources to a Container.

{% include feature-state-alpha.md %}
{% include feature-state-deprecated.md %}

{% endcapture %}

Expand Down