Following the instructions for GKE it uses the latest version of K8S by default. You should be able to deploy a full working Knative install (serving, build, eventing with logging, trace etc )
I followed the instructions, and after a bit of debugging, I went to Kabana and realised that it was complaining about non-matching indexes for logstash-* This is because fluent-d deamon set failes to start due to these issues.
https://github.com/knative/serving/issues/2218
https://github.com/kubernetes/kubernetes/pull/65643
Would be great if someone could verify. I am just doing a test on 1.10.x but if this is the case, perhaps we need to temporarily update the docs while the namespace issues are resolved?
Confirmed that installing version 1.10.9-gke.5 on GKE I get a working DeamonSet.
kubectl get DaemonSet -n knative-monitoring
NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE
fluentd-ds 3 3 3 3 3 beta.kubernetes.io/fluentd-ds-ready=true 6m
@mdemirhan is there a minimum version of k8s that we need to state as required to avoid the fluentd daemonset issue?
This is an issue with the newer versions. I will send a PR for a fix shortly for this one.
The PR will be for the head. Do you want to have a patch for the current release as well?
The PR will be for the head. Do you want to have a patch for the current release as well?
HEAD in the docs repo reflects the current release. If this is a change for v0.2.next, we should not merge until thats available.
Should we at least update the docs to state if you want logging, don't use K8S 1.11.x until we have a fix to save anyone doing what I did and scratching my head as to why it wasn't working?
@mdemirhan has this been resolved?
Yes, this is resolved.