Story (Required)
As a QE I want to run end-to-end tests as quickly as possible. Tests for s2i tasks run relatively slow.
Background (Required)
Tests for cluster tasks s2i-* get the list of image streams for particular runtime in namespace "openshift" and then run the test with all tags, e.g. for java on OpenShift 4.15 it would run with these tags/images
- java:11 registry.redhat.io/openjdk/openjdk-11-rhel7:latest
- java:8 registry.redhat.io/redhat-openjdk-18/openjdk18-openshift:latest
- java:openjdk-11-el7 registry.redhat.io/openjdk/openjdk-11-rhel7:latest
- java:openjdk-11-ubi8 registry.redhat.io/ubi8/openjdk-11:latest
- java:openjdk-17-ubi8 registry.redhat.io/ubi8/openjdk-17:latest
- java:openjdk-8-el7 registry.redhat.io/redhat-openjdk-18/openjdk18-openshift:latest
- java:openjdk-8-ubi8 registry.redhat.io/ubi8/openjdk-8:latest
- java:latest openjdk-17-ubi8 (this one would be omited)
It runs in step "Start and verify pipeline..." or "Start and verify dotnet pipeline..." in https://github.com/openshift-pipelines/release-tests/blob/master/steps/cli/tkn.go and it runs sequentially.
Approach (Required)
Run s2i tests in parallel
Notes
The same should be done for ecosystem tasks and step actions (probably part of 1.16)
- is related to
-
SRVKP-5552 [Downstream CI] Create automated tests for ecosystem tasks
-
- To Do
-