https://rancher.com/ logo
#harvester
Title
# harvester
a

adamant-kite-43734

12/28/2022, 9:53 PM
This message was deleted.
b

bland-farmer-13503

12/29/2022, 3:19 AM
May you try to modify
spec.replica
from 0 to 1 in
Copy code
kubectl edit -n cattle-monitoring-system prometheus rancher-monitoring-prometheus
Thanks.
b

best-hair-19862

12/29/2022, 12:30 PM
Thanks for the advice. I couldn't try it, in the end I reinstalled Harvester, couldn't wait 😅 . But I take note on that for future cases. Now I'm dealing with CPU constraints: I'm aware that I'm testing Harvester in a way lower-spec than the minimum requirements (24GB of RAM but only one quad-core CPU without hyperthreading...), so the monitoring pods aren't able to complete its initialization. I've tried to reduce the required CPU for Prometeus and Grafana to only 50m and I'd want to restart their deployment to see if there's any progress. Any advice on where could I lower the CPU reservations or is it useless to keep trying with this hardware?
So after some additional tweaking dropping StorageClass Replica counts to 1 and restarting the server several times, now I have all the monitoring services up and running and can access both Prometheus and Grafana, but
rancher-monitoring
app appears as Failed. Any thoughts? According to the Release Notes of the application, all related pods are up and running:
Copy code
# Run kubectl commands inside here
# e.g. kubectl get all
> kubectl --namespace cattle-monitoring-system get pods -l "release=rancher-monitoring"
NAME                                                     READY   STATUS    RESTARTS        AGE
rancher-monitoring-operator-bd988f5fb-jzh9h              1/1     Running   3 (4h28m ago)   21h
rancher-monitoring-prometheus-adapter-79d8db9697-h8bp8   1/1     Running   3 (4h28m ago)   21h
rancher-monitoring-prometheus-node-exporter-f5rbq        1/1     Running   2 (4h28m ago)   8h
2 Views