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

add election controlled service endpoints #1

Merged
merged 1 commit into from
Aug 20, 2020

Conversation

WanzenBug
Copy link
Contributor

Manually setting the endpoints for a service ensures that only the
elected pods receive traffic. This works around issues where controlling
traffic via pod readiness works suboptimal (i.e. rolling deployments).

In rolling deployments, readiness of a pod controls the further roll-out
of new versions. This means using readiness to control the service proxy
would lead to a deadlock: the new pods would never become "ready" as an old
pod still holds the lease, but the old pod is never terminated, as the rolling
update only continues if the new pods are ready.

@WanzenBug WanzenBug requested a review from JoelColledge August 20, 2020 08:14
@WanzenBug
Copy link
Contributor Author

This feature is motivated by the current effort to provide multiple replicas for the linstor-controller.

See also this comment describing the issues encountered with v0.1.0

Copy link
Member

@JoelColledge JoelColledge left a comment

Choose a reason for hiding this comment

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

This is a clever solution to an annoying problem 🙂

README.md Outdated Show resolved Hide resolved
README.md Outdated Show resolved Hide resolved
Manually setting the endpoints for a service ensures that only the
elected pods receive traffic. This works around issues where controlling
traffic via pod readiness works suboptimal (i.e. rolling deployments).

In rolling deployments, readiness of a pod controls the further roll-out
of new versions. This means using readiness to control the service proxy
would lead to a deadlock: the new pods would never become "ready" as an old
pod still holds the lease, but the old pod is never terminated, as the rolling
update only continues if the new pods are ready.
@JoelColledge JoelColledge merged commit ddeca69 into LINBIT:master Aug 20, 2020
@WanzenBug WanzenBug deleted the endpoint-control branch August 20, 2020 09:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants