When installing tiller
the server-side component of helm
in kubernetes,
with the default cri-o
trust level as untrusted
using kata-runtime,
the connection to tiller
pod errors out as shown below.
E0511 00:04:03.578489 10313 portforward.go:331] an error occurred forwarding 44249 -> 44134: error forwarding port 44134 to pod dcb2b2ed780469470e4fe1ec085fa02efc492516a79b6b1ce6bb90997997fdc7, uid : exit status 1: 2018/05/11 00:04:03 socat[11733] E connect(5, AF=2 127.0.0.1:44134, 16): Connection refused
Error: could not ping Tiller: rpc error: code = Unavailable desc = transport is closing
Current work-around is to set the environment variable HELM_HOST
to point to the clusterIP
of the tiller-deploy
service.
export HELM_HOST="$(kubectl -n kube-system get svc | grep tiller-deploy | awk '{print $3}'):44134"