[GEP-28] Do not watch Machines in Worker controller
#11971
Conversation
…autonomous shoots
Note the restarts:
```
root@machine-0:/# k get po -A | grep provider-local
extension-provider-local gardener-extension-provider-local-56d999bb56-l4kv5 1/1 Running 17 (5m30s ago) 110m
extension-provider-local gardener-extension-provider-local-56d999bb56-sxfcr 0/1 CrashLoopBackOff 16 (3m6s ago) 110m
```
Logs:
```
{"level":"error","ts":"2025-04-22T11:11:43.069Z","msg":"Error executing the main controller command","error":"error running manager: failed to wait for worker caches to sync kind source: *v1alpha1.Machine: timed out waiting for cache to be synced for Kind *v1alpha1.Machine","stacktrace":"main.main\n\tgithub.com/gardener/gardener/cmd/gardener-extension-provider-local/main.go:24\nruntime.main\n\truntime/proc.go:283"}
```
For ASCs, we can check if the API is present. If yes, we are in the
medium-touch case, i.e., we can watch `Machine`s. If no, we are in the
high-touch case, i.e., we cannot watch them. This works because CRDs are
always applied before extensions are deployed in the ASC scenario.
|
@rfranzke: GitHub didn't allow me to request PR reviews from the following users: maboehm. Note that only gardener members and repo collaborators can review this PR, and authors cannot review their own PRs. In response to this:
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-sigs/prow repository. |
|
LGTM label has been added. Git tree hash: 9d806d2145798cd63b0822b0983c71ef5a4e0ae4
|
adec676 to
eebc61e
Compare
|
LGTM label has been added. Git tree hash: 15885508fe8735f7090246f011005663c5f07091
|
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: timebertt 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 |
How to categorize this PR?
/area ipcei
/kind bug
What this PR does / why we need it:
Currently,
provider-localgoes into aCrashLoopBackoffafter some time in ASCs. Note the restarts:Logs:
This is because the
MachineAPI is not present (at least for the high-touch scenario).With this PR, we can check if the
MachineAPI is present in case we are reconciling an ASC. If yes, we are in the medium-touch case, i.e., we can watchMachines. If no, we are in the high-touch case, i.e., we cannot watch them. This works because CRDs are always applied before extensions are deployed in the ASC scenario.Which issue(s) this PR fixes:
Part of #2906
Special notes for your reviewer:
/cc @timebertt @maboehm
Release note: