✨ Update controller runtime and cluster-api to the newest version (CAPI 1.10)

[guettli]

Based on PR of @cgroschupp 🌱 update controller runtime and cluster-api to newest version by cgroschupp · Pull Request #1623 · syself/cluster-api-provider-hetzner

closes #1662

Changes

Use separate struct for webhooks, and do not export them. Example:

- var _ webhook.Defaulter = &HCloudRemediation{}
+ var _ webhook.CustomDefaulter = &hcloudRemediationWebhook{}

---

Followed the capi deprecation:

// Deprecated: This predicate is deprecated and will be removed in a future version,
// use ClusterPausedTransitionsOrInfrastructureReady instead.

---

Use docker image of Lychee (do not install it via alpine)

FROM docker.io/lycheeverse/lychee:sha-3592972-alpine@sha256:6609738da2ce125c64658728bd6af93da45852e0744e3335e90c5a5719e080c1 as lychee

---

CloudInitStatus: Do now hide the underlaying error, when calling cloud-init status failed.

There are three cases:

• command was never called (for example host/port not reachable)
• command was called, but exit was non-zero
• command was called and exit was zero. Only in this case it makes sense that the code analyzes the output of the command.

Caller of CloudInitStatus() uses output.ExitStatus() now.

---

GetCloudInitOutput: Same as for CloudInitStatus: Use output.ExitStatus() to differentiate between error cases.

---

obtainHardwareDetailsNics: Fail, if no IP was found.

---

Avoid && in shell commands.

Related: guettli/bash-strict-mode: Bash Strict Mode

Example:

apt-get update && apt-get install -y cloud-init apparmor apparmor-utils

Was changed to:

apt-get update
apt-get install -y cloud-init apparmor apparmor-utils

BTW, it took some time to understand why cloud-init was not installed. The && made the failing apt-get update not terminate the script. Additionally, the error output was swallowed. That was the reason for refactoring to use output.ExitStatus().

---

fix: input.ClusterctlConfigPath is required for DumpAllResources

---

fix: input.KubeConfigPath is required for DumpAllResources

---

fix: action "ensure-provisioned" failed: failed to get cloud init output: %!w()' (show real error message now)

---

updated cilium

---

Regenerated the cillium and ccm e2e templates

---

logStatusContinuously():

Show conditions of the wl-clusters.

This is very helpful, as you can see while the e2e test is running, if an error occurs.

---

[janiskemper]

I cannot review the changes in the scripts though

[guettli]

I cannot review the changes in the scripts though

I am unsure what you mean. I only see && updated, so that there are two lines now. This intentional.

[guettli]

@cgroschupp thank you for your initial work. The PR is now merged. We will do a new release, when #1647 got merged.