diff options
| author |   Bryan Sullivan <bryan.sullivan@att.com> | 2017-12-10 12:10:31 -0800 |
|---|---|---|
| committer | Bryan Sullivan <bryan.sullivan@att.com> | 2017-12-10 13:10:18 -0800 |
| commit | 00a470b032fc94b7d340770928ed003099e41a31 (patch) | |
| tree | a15d1d069bdfb1d030dfafc706b5aab24a96693e /tools/kubernetes/k8s-cluster.sh | |
| parent | c5ad6561884f1ee8e868b7922c8b27bda588d6d9 (diff) | |
AIO and centos deploy support
JIRA: MODELS-2
serial worker deploy
ceph not yet working for multi-node centos
Change-Id: Id46e3eb10486fd96b4acab45c87be8475a1a7163
Signed-off-by: Bryan Sullivan <bryan.sullivan@att.com>
Diffstat (limited to 'tools/kubernetes/k8s-cluster.sh')
| -rw-r--r-- | tools/kubernetes/k8s-cluster.sh | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/tools/kubernetes/k8s-cluster.sh b/tools/kubernetes/k8s-cluster.sh index f51b169..cac35e3 100644 --- a/tools/kubernetes/k8s-cluster.sh +++ b/tools/kubernetes/k8s-cluster.sh @@ -232,6 +232,9 @@ sudo kubeadm reset sudo $k8s_joincmd EOF +# process below is serial for now; when workers are deployed in parallel, +# sometimes calico seems to be incompletely setup at some workers. symptoms +# similar to as noted for the "wait for calico" steps above. for worker in $workers; do host=$(ssh -x -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no $USER@$worker hostname) log "Install worker at $worker hostname $host" @@ -247,10 +250,7 @@ EOF start_worker.sh $USER@$worker:/home/$USER/. ssh -x -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no \ $USER@$worker bash start_worker.sh - done - for worker in $workers; do - host=$(ssh -x -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no $USER@$worker hostname) log "checking that node $host is 'Ready'" status=$(kubectl get nodes | awk "/$host/ {print \$2}") while [[ "$status" != "Ready" ]]; do |