I have deployed istio with the following configuration:
helm install install/kubernetes/helm/istio --name istio --namespace istio-system \
--set global.controlPlaneSecurityEnabled=true \
--set mixer.adapters.useAdapterCRDs=false \
--set grafana.enabled=true --set grafana.security.enabled=true \
--set tracing.enabled=false \
--set kiali.enabled=false \
--set pilot.resources.requests.memory="264Mi" \
--set pilot.resources.requests.cpu="100m"Error:
istio-pilot-...-npv4m
Pending
0
a minute
0/1 nodes are available: 1 Insufficient cpu.Since this is a test env. I don't use a big sku (2 vcores and 4gb). But it looks that does not work, and this should not eat a lot of memory. Any help how to configure this will be appreciated.
This means that your CPU is already oversubscribed and it so happened that pilot doesn't have enough resources to start. What you can do is either use bigger\more nodes or remove CPU requests from the istio pods (most of the istio pods have quite substantial CPU requests). You might also want to check other pods (not istio), they might have CPU requests as well.