kubernetes: Ubuntu support broken on master (and likely several prior releases)

I was mostly following this guide, which seems to be either deprecated and/or broken: https://github.com/kubernetes/kubernetes/blob/master/docs/getting-started-guides/ubuntu.md

One key bug, one of the scripts references the wrong Github account, fix:

diff --git a/cluster/ubuntu/download-release.sh b/cluster/ubuntu/download-release.sh
index 3fce5d0..5fb3ea1 100755
--- a/cluster/ubuntu/download-release.sh
+++ b/cluster/ubuntu/download-release.sh
@@ -56,7 +56,7 @@ grep -q "^${ETCD_VERSION}\$" binaries/.etcd 2>/dev/null || {
 KUBE_VERSION=${KUBE_VERSION:-"1.1.2"}
 echo "Prepare kubernetes ${KUBE_VERSION} release ..."
 grep -q "^${KUBE_VERSION}\$" binaries/.kubernetes 2>/dev/null || {
-  curl -L https://github.com/GoogleCloudPlatform/kubernetes/releases/download/v${KUBE_VERSION}/kubernetes.tar.gz -o kubernetes.tar.gz
+  curl -L https://github.com/kubernetes/kubernetes/releases/download/v${KUBE_VERSION}/kubernetes.tar.gz -o kubernetes.tar.gz
   tar xzf kubernetes.tar.gz
   pushd kubernetes/server
   tar xzf kubernetes-server-linux-amd64.tar.gz

But I still can’t kube-up.sh:

$ export KUBE_VERSION=1.1.2
$ export KUBERNETES_PROVIDER=ubuntu

$ export nodes="root@10.0.10.xxx root@10.0.10.yyy"
$ export roles="ai i"
$ export NUM_MINIONS=2
$ export PATH=$PATH:~pwais/kubernetes/cluster/ubuntu/binaries

$ ./cluster/kube-up.sh 
... Starting cluster using provider: ubuntu
... calling verify-prereqs
... calling kube-up
/home/pwais/kubernetes/cluster/ubuntu /home/pwais/kubernetes
Prepare flannel 0.5.0 release ...
Prepare etcd 2.2.0 release ...
Prepare kubernetes 1.1.2 release ...
Done! All your binaries locate in kubernetes/cluster/ubuntu/binaries directory
/home/pwais/kubernetes

Deploying master and node on machine 10.0.10.240
saltbase/salt/generate-cert/make-ca-cert.sh: No such file or directory
config-default.sh                                                                                                                                100% 3441     3.4KB/s   00:00    
ubuntu/util.sh: No such file or directory
ubuntu/minion/*: No such file or directory
ubuntu/master/*: No such file or directory
ubuntu/reconfDocker.sh: No such file or directory
ubuntu/binaries/master: No such file or directory
ubuntu/binaries/minion: No such file or directory

I had more luck with v1.0.0 except that kube-proxy failed to launch on my minion because of an incorrect flag to the executable 😛

If somebody can direct me to a bona fide Ubuntu bare-metal setup guide, please LMK. Otherwise I guess I’ll try to poke at master a bit more (and perhaps contribute a patch if I get anywhere).

About this issue

Original URL
State: closed
Created 9 years ago
Comments: 38 (22 by maintainers)

Most upvoted comments

The line -listen-client-urls -listen-client-urls http://127.0.0.1:4001,http://${1}:4001\ in util.sh is problematic and should be changed to -listen-client-urls http://${1}:4001\ . it is causing the configuration in /etc/default/etcd configuration to have duplicate etc client url entry.

kostasp on Apr 19, 2016

I’ve run into the same issue, removing the http://${1}:4001 change something but the deployment still fail with the following outout. It continue retry but never did.

flanneld.conf                                                                                                                      100%  570     0.6KB/s   00:00
flanneld                                                                                                                           100% 2131     2.1KB/s   00:00
etcd start/running, process 3316
Error:  501: All the given peers are not reachable (Tried to connect to each peer twice and failed) [0]
Error:  501: All the given peers are not reachable (Tried to connect to each peer twice and failed) [0]
Error:  100: Key not found (/coreos.com) [16]
{"Network":"172.16.0.0/16", "Backend": {"Type": "vxlan"}}
{"Network":"172.16.0.0/16", "Backend": {"Type": "vxlan"}}
sudo: brctl: command not found
stop: Unknown instance:
docker start/running, process 3499
Connection to 127.0.0.1 closed.
Validating master
Validating root@127.0.0.1
Using master 127.0.0.1
cluster "ubuntu" set.
user "ubuntu" set.
context "ubuntu" set.
switched to context "ubuntu".
Wrote config for ubuntu to /home/ubuntu/.kube/config
... calling validate-cluster
Waiting for 1 ready nodes. 0 ready nodes, 1 registered. Retrying.
Waiting for 1 ready nodes. 0 ready nodes, 1 registered. Retrying.
Waiting for 1 ready nodes. 0 ready nodes, 1 registered. Retrying.

ido-ran on Aug 18, 2016