Added tutorials for features

[k82cn]

DESCRIPTION: -- symptom of the problem a customer would see

For now, kube-batch supports several features as follow; but we only have basic tutorial and design doc for gang-scheduling. It's better to have related docs for all features.

• Gang-scheduling/Coscheduling (KEP: Coscheduling kubernetes/community#2337)
• Preemption/Reclaim
• Task Priority within Job
• Queue
• DRF for Job sharing within Queue

[Jeffwan]

I can help on this issue and I think that's a great opportunity to learn this project. My plan is to try all these features and then write tutorials.

[k82cn]

that's great :)
I'll remove help-wanted label for you.

[Jeffwan]

/assign @Jeffwan

[afortiorama]

Hi,

I'm trying to understand how kube-batch would work? I assume scheduling based on resources is left to kubernetes scheduler (at least this is what I understand from your kube-batch image. Would you schedule jobs according to priorities and user fairshares like a classic batch system? I'm afraid it is not clear from the documentation. If not (yet?) is it something you are planning?

[k82cn]

I assume scheduling based on resources is left to kubernetes scheduler

kube-batch is another scheduler here; it will make the scheduling decision directly.

Would you schedule jobs according to priorities and user fairshares like a classic batch system?

yes, that's what're doing here.

[afortiorama]

Another maybe silly question. "DRF for Job sharing within Queue" does it mean you are going to use Django for authentication and authorization on use of resources?

[afortiorama]

That's another DRF I guess. But it would be interesting to know if you have any authentication/authorization policy layer.

[k82cn]

DRF

it's only one of policy, there're some others policy/algorithm here.

any authentication/authorization policy layer

It's based on current k8s feature; we did not handle that in scheduler.

[zoux86]

Hi,
where to find these detail tutorials and design doc for Queue or gang-scheduling.
thanks.

[Jeffwan]

@zoux86 I think that's the only tutorial now. https://github.com/kubernetes-sigs/kube-batch/blob/master/doc/usage/tutorial.md

Check a real example here. https://github.com/kubeflow/website/blob/master/content/docs/other-guides/job-scheduling.md

Queue part is missing in the tutorial. Let me finish tutorial this week.

[yylin1]

@Jeffwan I look forward to seeing your tutorial .
I am following kube-batch on kubeflow in my research.

[zoux86]

@Jeffwan is there a documentation about Queue.
Do you know what role the Queue playing in. #670
i will be very appreciated if you can help me to solve these confusion.Thanks.

[Jeffwan]

@zoux86 sorry I was a little bit busy with some releases this week. Already kick off this task and add more uses cases in the doc.

[k82cn]

/reopen

Let's keep this open until we have doc for each feature :)

[k8s-ci-robot]

@k82cn: Reopened this issue.

In response to this:

/reopen

Let's keep this open until we have doc for each feature :)

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[k82cn]

/cc @asifdxtreme

[asifdxtreme]

#703

[yylin1]

Hi, In tutorials could add more about the scheduler configuration ?
Such as the plugin configuration and detailed plugin arguments example about the ConfigMap.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.