-
Bug
-
Resolution: Done
-
Critical
-
Pipelines 1.18.0
-
3
-
False
-
-
False
-
Fixes Result config options propagation from TektonConfig CR to TektonResults object.
-
-
-
1
-
Pipelines Sprint Crookshank 26, Pipelines Sprint Crookshank 27, Pipelines Sprint Crookshank 28
Description of problem:
Using the TektonConfig CR does not propagate config options for Results to the TektonResults object. This prevents configuration of Tekton Results: TektonConfig never writes the changes, and prevents modifications from being made to the TektonResults CR as well.
Workaround: It should be possible to modify the TektonInstallerSet for results instead (results-##### on any given install). I believe this may persist until another version is installed.
Prerequisites (if any, like setup, operators/versions):
- OpenShift Pipelines 1.18.0
Steps to Reproduce
# (Tested with OCP 4.16.37 on OCP Pipelines 1.18.0)
- Install OpenShift Pipelines 1.18.0 to an OpenShift Cluster
- Attempt to modify TektonConfig with changes such that
spec:
results:
options:
deployments:
tekton-results-watcher:
metadata:
creationTimestamp: null
spec:
selector: null
strategy: {}
template:
metadata:
creationTimestamp: null
spec:
containers:
- args:
- --store_event=true
- --store_deadline=10s
- --forward_buffer=10s
name: watcher
resources: {}
status: {}
Actual results:
No changes are made.
Expected results:
TektonConfig pushes changes to TektonResults, which changes the affected Deployment, tekton-results-watcher.
Reproducibility (Always/Intermittent/Only Once):
Always
Acceptance criteria:
TektonConfig correctly propagates changes to subordinate resources.