search

techno-tim / k3s-ansible

The easiest way to bootstrap a self-hosted High Availability Kubernetes cluster. A fully automated HA k3s etcd install with kube-vip, MetalLB, and more. Build. Destroy. Repeat.
https://technotim.live/posts/k3s-etcd-ansible/
Apache License 2.0
2.41k stars 1.05k forks source link

Metallb webhook-service renamed to metallb-webhook-service #487

Closed black-dragon74 closed 6 months ago

black-dragon74 commented 7 months ago

While deploying the manifests for MetalLB version v0.14.4, the wehbhook-service is named metallb-webhook-service.

This causes the post task to fail with an error.

Posting it here for anyone else facing the same. I can send a PR for the same.

It would make sense to dynamically set a var for service name depending of the config values: metal_lb_*_tag_version.

WDYT @timothystewart6 ?

Regards

timothystewart6 commented 7 months ago

Hey! I am confused on how this is happening? Is this in an upgrade scenario? Or using an old tag with the latest playbook?

tfs commented 7 months ago

Hey! I am confused on how this is happening? Is this in an upgrade scenario? Or using an old tag with the latest playbook?

Happened to me, too. Scenario for me: changed the version for metallb in the all.yml in my inventory. Redeployed the cluster by running the playbook.

black-dragon74 commented 7 months ago

Hey! I am confused on how this is happening? Is this in an upgrade scenario? Or using an old tag with the latest playbook?

They changed the webhook service name starting with v0.14.4 (added metallb prefix) to more accurately denote where the service comes from.

timothystewart6 commented 7 months ago

Got it. Hmmm.... Not sure I want to make any changes since this is a point in time, otherwise we would have to account for every breaking change with every included dependency. We always ensure that the current configuration works on clean machines (since we run it in CI). Generally speaking (although it does work) upgrading using the playbook is kind of an edge case since we recommend using helm or gitops to handle this after the cluster is bootstrapped.

If you're saying that it's broken in the latest version of Metal LB, happy to review a PR for it!

konstantin-kornienko commented 4 months ago

Trying to fix this issue, faced it too. https://github.com/techno-tim/k3s-ansible/pull/528/files