Consider exposing petset index via downward api

[bprashanth]

While it's easy enough to get through something like:
IFS='-' read -ra LIST <<< "$POD_NAME"; INDEX=${LIST[1]}
It requires a custom entrypoint. If the ordinal index is a common part of 12 fa apps we should just expose it more naturally through the downward api.

ref #30256
@tobegit3hub

[nhlfr]

I'm going to work on this.

[bprashanth]

@kubernetes/api-review-team @pmorie who to tag for downward api requests?
@smarterclayton out of petset interest

[smarterclayton]

I agree that this is useful, although we need to be careful about how to
expose it.

On Wed, Aug 17, 2016 at 4:11 PM, Prashanth B notifications@github.com
wrote:

@kubernetes/api-review-team
https://github.com/orgs/kubernetes/teams/api-review-team @pmorie
https://github.com/pmorie who to tag for downward api requests?
@smarterclayton https://github.com/smarterclayton out of petset interest

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#30427 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABG_p9eA899Ppwfik9_u1R2vCJGusr29ks5qg2sIgaJpZM4JhzXh
.

[nhlfr]

I'd suggest to keep a petset index in pod annotation (i.e. pod.beta.kubernetes.io/petset) and then pass it through downward API.

[smarterclayton]

It would need to be available via environment as well.

[nhlfr]

Of course. What I'm proposing is:

• applying pod.beta.kubernetes.io/petset-index annotation with pet set index to any pods created during pets creation
• adding dependencies.petSetIndex label which may be used for environment downward API
• using the info from pod.beta.kubernetes.io/petset-index annotation to fill up the env with dependencies.petSetIndex field reference

[nhlfr]

Due to discussion in PR, I decided to make an another issue about exposing annotations via downward API - #31218

In my PR, which is referring to this issue, I will only put an annotation with petset index on dependent pods. Then the general downward API for all annotations will make this petset index accessible.

I decided to use annotations, because I (and probably @smarterclayton ) think that we shoud'nt add any fields to Pod in case of alpha/beta features like petsets.

[alexouzounis]

is there a workaround for this currently ? If we don't want to change the entry point that is.

[smarterclayton]

You can expose pod name via the downward API.

On Mon, Oct 31, 2016 at 1:08 PM, Alex Ouzounis notifications@github.com
wrote:

is there a workaround for this currently ? If we don't want to change the
entry point that is.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#30427 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABG_pxPnQufQk_29a18fSKcaKgcgMBiaks5q5iCQgaJpZM4JhzXh
.

[bprashanth]

#30427 (comment), you can even replace POD_NAME with $(hostname) if you're using statefulset (aka petset)

[alexouzounis]

yeah - we are using PetSet and what we want is just the ordinal exposed in an environment variable i.e. 0, 1 etc. The pod name will be something like web-0 and i can't find a way to expose that in an env var.

[nhlfr]

Sorry for leaving this issue without any action. PTAL on the issue #31218. It will surely not be touched in 1.5 release, but I've submitted a proposal to be considered for a future release.

[jberkus]

+1 on this feature, regardless of how it's implemented. Currently I'm parsing the pod name, which is awkward and fragile.

[alexouzounis]

Indeed - we ended doing the same as a postStart lifecycle hook to avoid modifying the docker image

[smarterclayton]

@pmorie as we discussed we need to fully specify the downward API for
complex fields (annotations and labels) and get that over the review hump
in 1.6 so people can use an annotation in an env var. @liggitt too maybe.

On Tue, Nov 15, 2016 at 12:58 PM, Alex Ouzounis notifications@github.com
wrote:

Indeed - we ended doing the same as a postStart lifecycle hook to avoid
modifying the docker image

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#30427 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABG_p9oxBplas5h6lnuQsiFNU2U3YSUYks5q-h0SgaJpZM4JhzXh
.

[Aisuko]

I tried the way name Pod hook, but they did not work for me. I want to get pods id of statefulsetl and add it to containers ENV. So, there is my solution below. It works for me.

env:
- name: POD_ID_OF_STATEFULSET 
   valueFrom:
     fieldRef:
       fieldPath: metadata.name

  POD_ID_OF_STATEFULSET:      phoenix-package-0 (v1:metadata.name)

[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

[jberkus]

This issue has an active PR.

/remove-lifecycle rotten

[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

[jlegrone]

/remove-lifecycle rotten

[jberkus]

/remove-lifecycle rotten

This issue has an active PR

[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

[bohdantverdyi]

/remove-lifecycle rotten

This issue has an active PR

[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.

[bohdantverdyi]

/remove-lifecycle rotten

This issue has an active PR