We have this project deployed on k8s and while it works it is generating a lot of logs (current count 130k log entries every hour). Observing the pod definitions this project seems to deploy fixed debug/trace log levels to its components. I may be able to set the operator log level, but the CSI_LOGLEVEL is fixed set on trace in
|
func (s *csiNodeSyncer) getEnvFor(name string) []corev1.EnvVar { |
and the controller is also fixed on
|
Value: config.DefaultLogLevel, |
(which may seem configurable but is a constant in settings.go). While doing this I noted that a LOT of settings on these deployments/daemonset are hardcoded. You can't even change them after the operator has deployed them, since the operator will just revert the changes.
If this issue is not fixed I'm seriously considering deploying the hostdefiner and controller deployments, and the csi-node daemonset manually, which would be a large waste of time of nothing else.
We have this project deployed on k8s and while it works it is generating a lot of logs (current count 130k log entries every hour). Observing the pod definitions this project seems to deploy fixed debug/trace log levels to its components. I may be able to set the operator log level, but the CSI_LOGLEVEL is fixed set on trace in
ibm-block-csi-operator/controllers/syncer/csi_node.go
Line 233 in 62f9c36
ibm-block-csi-operator/controllers/syncer/csi_controller.go
Line 333 in 62f9c36
If this issue is not fixed I'm seriously considering deploying the hostdefiner and controller deployments, and the csi-node daemonset manually, which would be a large waste of time of nothing else.