workload-cluster: robustness improvement for dummy units (!8581) · Merge requests · Sylva-projects / sylva-core

Context: we've seen platforms that for historical reasons still were using units.sriov.enabled: true in workload clusters; this unit is a dummy unit (producing no resource), but workload context defaults give it a spec.kubeConfig to point to the workload cluster... as a result if the unit reconciles before the kubeconfig secret is created, things break.

We don't encounter this problem in CI because in workload-cluster.values.yaml the dummy units we use inherit from kustomization-deployed-on-mgmt-cluster (which clears the kubeConfig). A possibility could have been to fix the sriov unit to also have it inherit from kustomization-deployed-on-mgmt-cluster, or clear its kubeConfig. This is what we'll do to improve release branches (see #4392 (closed)).

But in fact, since dummy units never need a spec.kubeConfig (because they don't create resources), a cleaner solution is to have the dummy unit template clear the kubeConfig upfront. This is point (A).

... later edit:

point (B): as pointed out below by @ishitamittal.2301, this is incorrect since we have some units inheriting from dummy which create resources
- for these the cleanest is to stop using the dummy unit template
point (C): I later realized that another use-case for dummy unit is healthchecks done in the remote cluster: these need the kubeConfig as well

What this MR is doing:

bring this fix (well, robustness improvement) which avoids having an irrelevant kubeConfig on dummy units
do a related code simplification (cluster-ready unit does not anymore need to inherit from kustomization-deployed-on-mgmt-cluster, so the unit_templates inherited from values.yaml does not need to be overloaded)
do a minor cleanup: cluster-reachable dummy unit unit_templates statement can be removed since it only repeats the one inherited from values.yaml (this was already true before this MR)
to address point (B), a separate commit adjusts some xxx-tls units in workload-cluster.values.yaml to avoid using the dummy unit template
to address point (C), a separate commit introduces a dummy-for-in-cluster-healthchecks, dedicated to units that today inherit from dummy to do healthcheck in the remote cluster -- this solves point (C) and also simplifies the code a bit by avoiding the repeat of the kubeConfig (which these units had to point to the remote cluster not only in workload context, but also in bootstrap context)

CI configuration

Below you can choose test deployment variants to run in this MR's CI.

Click to open to CI configuration

Legend:

Icon	Meaning	Available values
☁️	Infra Provider	`capd`, `capo`, `capm3`
🚀	Bootstrap Provider	`kubeadm` (alias `kadm`), `rke2`, `okd`, `ck8s`
🐧	Node OS	`ubuntu`, `suse`, `na`, `leapmicro`
🛠️	Deployment Options	Deployment option list and description
🎬	Pipeline Scenarios	Available scenario list and description
🟢	Enabled units	Any available units name, by default apply to management and workload cluster. Can be prefixed by `mgmt:` or `wkld:` to be applied only to a specific cluster type
🔴	Disabled units	Any available units name, by default apply to management and workload cluster. Can be prefixed by `mgmt:` or `wkld:` to be applied only to a specific cluster type
🏗️	Target platform	Can be used to select specific deployment environment Available platform list and description
⚡	Pipeline control	`autorun`, `manual` or `blocking`. Can be used to override global config and start a deployment pipeline the required way

Global config for deployment pipelines

autorun pipelines
allow failure on pipelines
record sylvactl events

Notes:

Enabling autorun will make deployment pipelines to be run automatically without human interaction
Disabling allow failure will make deployment pipelines mandatory for pipeline success.
if both autorun and allow failure are disabled, deployment pipelines will need manual triggering but will be blocking the pipeline

Be aware: after configuration change, pipeline is not triggered automatically. Please run it manually (by clicking the run pipeline button in Pipelines tab) or push new code.

Edited Sep 03, 2026 by Thomas Morin

workload-cluster: robustness improvement for dummy units

CI configuration

Global config for deployment pipelines

Merge request reports