-
Notifications
You must be signed in to change notification settings - Fork 224
Update Kubernetes from v1.11.3 to v1.12.1 #1003
Update Kubernetes from v1.11.3 to v1.12.1 #1003
Conversation
Can one of the admins verify this patch? |
1 similar comment
Can one of the admins verify this patch? |
Can one of the admins verify this patch? |
These changes overall seem fine to me. On a quick scan of the issue, however, I'm not clear on how these certs are used. Is this something we should open an issue for and come back and revisit minting real certs for? Then also - we need the tests to be running before actually merging this. @rphillips was looking into this. |
ok to test |
Yeah sorry, it was mainly intended to repro the issue. The
controller-manager generates certs and I'm thinking we may want to mint
certs
…On Thu, Oct 11, 2018, 4:05 PM Aaron Levy ***@***.***> wrote:
ok to test
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#1003 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ACJidNKT9vlcbOAW49qWN9iggsh4y7SDks5uj864gaJpZM4XXneD>
.
|
Can one of the admins verify this patch? |
1 similar comment
Can one of the admins verify this patch? |
I'll rebase to pickup the new pod-checkpointer image once #1010 lands. |
* Workaround kubernetes/kubernetes#68973 by providing an emptyDir for auto-generated certs
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: If they are not already assigned, you can assign the PR to them by writing The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
Concerning the controller-manager generated certs, I poked around and asked some questions. kubernetes/kubernetes#68973 (comment) I kinda favor not bothering to mint real certs for these yet. The HTTPS endpoint seems to be a new and so far unused endpoint that doesn't have clients. The HTTP endpoint is still always there. Nothing even happens if the HTTPS endpoint is turned off (--secure-port=0). Seems half baked / incomplete. If we were aiming to tell users they can now talk to the controller-manager with TLS, I think we'd need to create a controller-manager service, use a conventional DNS name, mint certs with that name, and also tell people to just ignore the HTTP endpoint for now. Not sure what the use case is here really, feels kinda fruitless. 🤷♂️ |
I think this is ready to go. Might be valuable to get on v1.12 before the new cycle of manifest improvements. |
Looks like we missed these. |
The checkpointer image doesn't necessarily need to update those source files to work with v1.12. |
/lgtm |
emptyDir
for auto-generated controller-manager certs