I imagine this has already been addressed (if so point me to the thread), but wh . . .

Kubernetes-Slack Discussions
#1

Jay Scruggs:
I imagine this has already been addressed (if so point me to the thread), but when doing the etcd snapshot restore the kube-api server seems to permanently lose the ability to connect to etcd. The kube-api container logs indicate an authentication problem.

  1 clientconn.go:1331] [core] grpc: addrConn.createTransport failed to connect to {127.0.0.1:2379 127.0.0.1 <nil> 0 <nil>}. Err: connection error: desc = "transport: Error while dialing dial tcp 127.0.0.1:2379: connect: connection refused". Reconnecting...
#2

dexterous neo:
yes, assuming you restore it to a new --data-dir location, you’d have to configure it back in ETCD and give it a minute. It will establish the connection then after

#3

Jay Scruggs:
Right, I did that but it never re-establishes a connection.

#4

Radoslaw Czajkowski:
is etcd running in a pod?

#5

Jay Scruggs:
I finally got it to work. But to answer your question, no it wasn’t running. Thanks for following up.

#6

dexterous neo:
@Jay Scruggs will you please share what was broken? and how did you fix it?

#7

Jay Scruggs:
I suspect it was because used localhost:2379 for the --endpoints value instead of 127.0.0.1:2379 . I wouldn’t expect that to matter but when I used the number format, it worked.

#8

dexterous neo:
@Jay Scruggs good