The current workload templates has ConfigMaps that embed all logic. This makes difficult to read and contribute to them.
For example this template https://github.com/openshift-scale/workloads/blob/master/workloads/templates/workload-mastervertical-script-cm.yml.j2 embeds run.sh, conprof_start.sh, conprof_stop.sh, workload.sh and all clusterloader templates.
With the current approach there're duplicated code (for example conprof_start.sh is present in other templates).
As an idea to avoid this we could use the command kubectl create cm --from-file=... to create on-fly the ConfigMaps from the scripts required by the workload.
The current workload templates has ConfigMaps that embed all logic. This makes difficult to read and contribute to them.
For example this template https://github.com/openshift-scale/workloads/blob/master/workloads/templates/workload-mastervertical-script-cm.yml.j2 embeds run.sh, conprof_start.sh, conprof_stop.sh, workload.sh and all clusterloader templates.
With the current approach there're duplicated code (for example conprof_start.sh is present in other templates).
As an idea to avoid this we could use the command
kubectl create cm --from-file=...to create on-fly the ConfigMaps from the scripts required by the workload.