Claus Ibsen created CAMEL-26934:
-----------------------------------
Summary: camel-jbang-plugin-kubernetes - --kube-config is ignored
by delete, logs loops on a finished pod, and traits are dropped for CronJob and
Knative
Key: CAMEL-26934
URL: https://issues.apache.org/jira/browse/CAMEL-26934
Project: Camel
Issue Type: Bug
Components: camel-jbang
Reporter: Claus Ibsen
Fix For: 4.24.0
Found by a scan of the Camel CLI on 2026-11-13, with the unpacked
4.24.0-SNAPSHOT camel-launcher on macOS (JDK 25). *(exp)* means reproduced by
experiment, *(read)* means traced in the code. Paths are relative to
{{dsl/camel-jbang/}}, and line numbers are from main at that time. Earlier
findings of this kind: CAMEL-26901, CAMEL-26911 to CAMEL-26914.
No real cluster was used: the calls went to an unreachable kubeconfig or a
local fake API server. Paths are under
{{camel-jbang-plugin-kubernetes/.../commands/kubernetes/}}.
h3. 1. --kube-config is ignored in three places (exp + read, high)
{{KubernetesDelete.java:51}} uses the static default client,
{{KubernetesHelper.discoverClusterType}} (147-190) uses the default client, and
the jkube deploy ({{KubernetesRun}} 668-797) never gets the kubeconfig.
{{delete --kube-config=<fake>}} with {{KUBECONFIG}} pointing elsewhere made 0
requests to the given cluster: a delete aimed at staging acts on the current
context.
h3. 2. The client timeouts are set after the client is built (exp)
{{KubernetesHelper.getKubernetesClient}} (85-91): with an unreachable API {{run
--output yaml}} takes 40.7 s, against 1.8 s with the timeouts preset. Move one
line.
h3. 3. kubernetes logs loops forever on a finished pod (exp)
{{isPodTerminated}} (153-161) checks the phase "Terminated", which does not
exist, and {{retryCount}} is reset for every printed line (137), so the cap
never triggers. A Succeeded pod's log was printed 6 times in 12 s. Also hits
CronJob pods and {{run --logs}}/{{--dev}}.
h3. 4. Options dropped for CronJob and Knative Service (exp)
EnvTrait, MountTrait, LabelTrait and AnnotationTrait apply only to the
Deployment: a cronjob export with {{--env}}, {{--config}}, {{--volume}} and
{{--annotation}} exits 0 and none is in {{cronjob.yml}}. A Knative Service
loses the env and the labels, including {{app.kubernetes.io/runtime}}.
h3. 5. Duplicate volume names (exp)
{{MountTrait}} names a volume {{sanitize(name)}} with no type prefix and cuts
at the first dot: {{--config configmap:db --config secret:db}} gives two
volumes named {{db}}, which the API server rejects.
h3. 6. Medium
* {{--profile}} drops the traits of {{application.properties}}
({{KubernetesExport.java:295-309}}).
* {{kubernetes run --property}} drops any value with a {{/}} as a file name
({{KubernetesRun.java:317-335}}), for example {{camel.server.path=/api}}.
* A comma in a trait value is split into a list even when quoted
({{TraitHelper.resolveTraitValue}} 182-188): {{cronjob.schedule=0 1,13 * * *}}
exits 70 with a Jackson trace.
* {{ingress.annotations=k=v}} exits 70; the doc example at
{{camel-jbang-kubernetes-advanced.adoc:378}} uses {{-t}}, an unknown option;
with {{--trait}} the annotation is ignored.
* Image settings: {{--registry-mirror}} builds
{{mirror/mirror.gcr.io/library/...}}; {{--image=localhost:5000/app:1}} gets no
registry or insecure setting; 4-part names get no registry; {{export
--cluster-type=openshift}} has no namespace in the image, unlike run.
* {{KubernetesHelper.sanitize}}: {{my.app.v2}} becomes {{my}}, {{2048-game}} is
not a valid DNS-1035 name, {{___}} gives an empty name and exit 0.
* {{cronjob.enabled=true}} without a schedule generates no workload and exits
0; {{cronjob.backoff-limit=0}} is dropped (Kubernetes uses 6).
h3. 7. Low
* A Service and port 8080 are always generated, also for timer routes and
CronJobs; the check is dead code and the port comment is stale (health is on
9876).
* {{ingress.auto}} is documented but never read.
* A custom container name is reported as "not supported" but applied.
* On Windows the mount path uses backslashes.
* {{delete}} never deletes CronJobs and exits 0 when nothing is found ({{logs}}
exits 3).
* {{run --output yaml}} still contacts the cluster unless {{--cluster-type}}
and {{--disable-auto}} are given; the detection message goes to {{System.out}}
({{KubernetesHelper.java:163,183}}) inside the YAML; {{--quiet}} swallows the
reasons of "Project export failed!"; typos in trait names are ignored;
{{delete}} and {{logs}} have no {{--json}}.
h3. Came out clean
{{--cluster-type}} validation, the kind and minikube registry mapping,
Deployment and Service consistency, the Ingress and Route mapping, {{--env}}
values with {{=}} or {{,}}.
_Claude Code on behalf of Claus Ibsen_
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)