[mw-0]

Description

This PR fixes #14244.

Since 4.23 the binaries ISO ships the Headlamp dashboard as headlamp.yaml instead of
dashboard.yaml. The create path (k8s-control-node.yml) already falls back between the two,
but upgrade-kubernetes.sh hardcoded dashboard.yaml. Because the script runs with bash -e,
every upgrade using a Headlamp ISO failed on the first control node after kubeadm had already
upgraded it, leaving the cluster half-upgraded and in Alert.

This applies the same fallback as the create path: headlamp.yaml, then dashboard.yaml,
then a warning, so ISOs built before and after the Headlamp change both work.

Affects 4.23.0.0 and main. 4.22 is not affected. Happy to retarget if a 4.23 branch is
wanted for a patch release.

Types of changes

• Bug fix (non-breaking change which fixes an issue)

Feature/Enhancement Scale or Bug Severity

Bug Severity

• Major

How Has This Been Tested?

• 4.23.0.0 management server with the patched script, single-control/single-worker CKS
  cluster, isolated network, .
• Upgraded 1.35.8 -> 1.36.0 using a Headlamp-built ISO: upgrade completes, both nodes on
  v1.36.0, headlamp running in kube-system, cluster back to Running, version updated in DB.
• Before the patch the same upgrade failed with
  error: the path "/mnt/k8sdisk//dashboard.yaml" does not exist.

How did you try to break this feature and the system with this change?

• tried to upgrade using custom isos
• Missing both files only logs a warning instead of failing the upgrade.### Description

[boring-cyborg]

Congratulations on your first Pull Request and welcome to the Apache CloudStack community! If you have any issues or are unsure about any anything please check our Contribution Guide (https://github.com/apache/cloudstack/blob/main/CONTRIBUTING.md)
Here are some useful points:

• In case of a new feature add useful documentation (raise doc PR at https://github.com/apache/cloudstack-documentation)
• Be patient and persistent. It might take some time to get a review or get the final approval from the committers.
• Pay attention to the quality of your code, ensure tests are passing and your PR doesn't have conflicts.
• Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Issues, Mailing list and Slack.
• Be sure to read the CloudStack Coding Conventions.
  Apache CloudStack is a community-driven project and together we are making it better 🚀.
  In case of doubts contact the developers at:
  Mailing List: dev@cloudstack.apache.org (https://cloudstack.apache.org/mailing-lists.html)
  Slack: https://apachecloudstack.slack.com/

[andrijapanicsb]

@ACSHomeBot package G

[ACSHomeBot]

Packaging results:

Result	Artifact	Platform
PASS	RPM	EL (EL8/9/10)
PASS	DEB	Ubuntu, Debian

Test packages are available at:

• EL9/RPM: https://f003.backblazeb2.com/file/andrijapanicsb-cloudstack-pr-builds/pr/14245/el9/
• Debian/APT: https://f003.backblazeb2.com/file/andrijapanicsb-cloudstack-pr-builds/pr/14245/debian/
• OpenPGP public key

[weizhouapache]

code lgtm

[mw-0]

extras removed from file

[Damans227]

can we drop the + at the start of these lines?

@mw-0 that covers it, the script parses fine now

[Damans227]

thanks @mw-0, fix looks right. LGTM

[nvazquez]

@blueorangutan package

[blueorangutan]

@nvazquez a [SL] Jenkins job has been kicked to build packages. It will be bundled with no SystemVM templates. I'll keep you posted as I make progress.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 20.39%. Comparing base (0a5bf30) to head (bcaf69a).
⚠️ Report is 81 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff              @@
##               main   #14245      +/-   ##
============================================
+ Coverage     20.05%   20.39%   +0.34%     
- Complexity    20656    21283     +627     
============================================
  Files          6427     6481      +54     
  Lines        584905   588040    +3135     
  Branches      71658    72021     +363     
============================================
+ Hits         117278   119951    +2673     
- Misses       454802   455058     +256     
- Partials      12825    13031     +206     

Flag	Coverage Δ
uitests	3.70% <ø> (-0.02%)	⬇️
unittests	21.70% <ø> (+0.37%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[blueorangutan]

Packaging result [SF]: ✔️ el8 ✔️ el9 ✔️ el10 ✔️ debian ✔️ suse15. SL-JID 19845

[andrijapanicsb]

@ACSHomeBot package

[ACSHomeBot]

Accepted: package kvm for bcaf69a5cb8d.

Planned: RPM + DEB, KVM SystemVM.

Status: Completed. Packaging results.

[ACSHomeBot]

Packaging results:

Result	Artifact	Platform	Build Time
PASS	RPM	EL (EL8/9/10)	40 min
PASS	DEB	Ubuntu, Debian	33 min

Test packages are available at:

• EL9/RPM: https://packages.v2kmigrate.com/cloudstack/pr/14245/el9/
• Debian/APT: https://packages.v2kmigrate.com/cloudstack/pr/14245/debian/

[bhouse-nexthop]

LGTM. The upgrade now picks the dashboard the same way k8s-control-node.yml does on create: headlamp.yaml first, then dashboard.yaml, then a warning. So ISOs built before and after #12776 both work.

This matters to us: our clusters were created on 4.19/4.20 with the legacy dashboard. Without this, their first Kubernetes upgrade with a 4.23+ ISO stops on the main control node after kubeadm and kubelet have already been upgraded, and the ISO stays mounted.

One non-blocking note, for a follow-up or the docs. A cluster created with the legacy dashboard and then upgraded with a Headlamp ISO ends up running both. The old kubernetes-dashboard deployment is never updated again, and the kubernetes-dashboard-ui cluster-admin binding that create gave its service account stays in place. The upgrade could remove the old namespace and binding once Headlamp is applied. Or the upgrade notes could tell operators to do it themselves.

[boring-cyborg]

Awesome work, congrats on your first merged pull request!