Closed anilkhichar closed 10 months ago
I'm not exactly sure what the goal of this issue is.
The recommended steps when installing the Solr Operator is to use kubectl install
and the steps when upgrading the Solr Operator use kubectl replace
.
The reason for this is that the last-applied-configuration can be too big for some Kubernetes instances, and those fields are useless for CRDs.
kubectl create/replace VS kubectl apply/apply
Can we get real data points specific to solr operator CRDs where using apply can get wrong. Based on my experiment there is no difference in the applied CRD content.
My experiment steps: Step1:
kubeclt get crd -o yaml > *.yaml
kubeclt get crd -o yaml > *.yaml
kubeclt get crd -o yaml > *.yaml
Step2:
kubeclt get crd -o yaml > *.yaml
kubeclt get crd -o yaml > *.yaml
kubeclt get crd -o yaml > *.yaml
Step3, Step4: Also repeated steps from v0.3.0 to v0.7.0 (but same result)
Now the generated outcome YAML files do have exact same content except one
last-applied-cofiguration
extra element in the apply. And we don't need to worry about it.Other references: https://github.com/apache/solr-operator/issues/502