apisix-helm-chart-experimental
Experimental project to make it easy to deploy apisix to k8s via helm, not finished yet!
NOTICE: This project is a experimental project, do not use it in production environment!!!
Install
# install by helm
NAMESPACE=apisix
helm install --name apisix --namespace $NAMESPACE .
ETCD cluster is created by etcd-operator helm chart.
Delete
-
Delete apisix chart
helm delete --purge apisix
-
Cleanup etcd CRDs (optional)
NAMESPACE=apisix kubectl get crd -n $NAMESPACE | grep coreos | awk '{ print $1 }' | xargs kubectl delete crd -n $NAMESPACE
Known issues
- When the upstreams are the services on k8s, we should use FQDN of
kube-dns
service as dns server address(kube-dns.kube-system.svc.cluster.local
), and theresolver
directive must havekube-dns
's FQDN only, otherwise nginx can not resolve the FQDN of service in k8s. Reference issue: #298
References
- Regarding ETCD backup and restore, see etcd backup operator &
etcd restore operator;
you can pass
--set etcd-operator.customResources.createBackupCRD=true
and--set etcd-operator.customResources.createRestoreCRD=true
and other values during install apisix by helm command.
TODO
- Work with istio.