Skip to content

Commit 9f88eef

Browse files
committed
Revise deployment instructions after deployment test
1 parent 844a500 commit 9f88eef

File tree

1 file changed

+13
-14
lines changed

1 file changed

+13
-14
lines changed

docs/deploy.md

Lines changed: 13 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -34,7 +34,7 @@ store your cluster's configuration and authorities.
3434

3535
Now, create an ISO:
3636

37-
$ spire iso gen preseeded.iso building/ ~/.ssh/id_rsa.pub
37+
$ spire iso gen preseeded.iso ~/.ssh/id_rsa.pub # this key is used for direct access during cluster setup
3838

3939
Now you should burn and/or upload preseeded.iso that you've just gotten, so
4040
that you can use it for installing servers. Make a note of the password it
@@ -73,10 +73,7 @@ For the official homeworld servers:
7373
* Configure the supervisor keyserver:
7474

7575
$ spire setup keyserver
76-
77-
* Check that the keyserver is running properly:
78-
79-
$ spire verify keystatics
76+
$ spire verify keystatics # make sure the keyserver is running
8077

8178
* Admit the supervisor node to the cluster:
8279

@@ -93,7 +90,6 @@ For the official homeworld servers:
9390

9491
$ spire access update-known-hosts # set up certificate authority in ~/.ssh/known_hosts
9592
$ spire access ssh # if this fails, you might need to make sure you don't have any stale kerberos tickets
96-
$ ssh-keygen -L -f ~/.ssh/id_rsa-cert.pub
9793

9894
* Configure and test SSH:
9995

@@ -142,10 +138,11 @@ For the official homeworld servers:
142138
member 439721bf885a52a5 is healthy: got healthy result from https://18.181.0.104:2379
143139
member 61712dffdce48432 is healthy: got healthy result from https://18.181.0.97:2379
144140
member f6d798ec325cf15d is healthy: got healthy result from https://18.181.0.106:2379
141+
cluster is healthy
145142

146143
* Query etcd cluster members:
147144

148-
$ spire etcdctl cluster-health member list
145+
$ spire etcdctl member list
149146
439721bf885a52a5: name=huevos-rancheros peerURLs=https://18.181.0.104:2380 clientURLs=https://18.181.0.104:2379 isLeader=false
150147
61712dffdce48432: name=eggs-benedict peerURLs=https://18.181.0.97:2380 clientURLs=https://18.181.0.97:2379 isLeader=true
151148
f6d798ec325cf15d: name=ole-miss peerURLs=https://18.181.0.106:2380 clientURLs=https://18.181.0.106:2379 isLeader=false
@@ -189,7 +186,8 @@ DNS, but when that happens, you can turn it back off:
189186

190187
## Bootstrap cluster registry
191188

192-
$ ln -s .../keys-for-homeworld.mit.edu/ $HOMEWORLD_DIR/https-certs
189+
$ mkdir $HOMEWORLD_DIR/https-certs
190+
$ cp homeworld.mit.edu.key homeworld.mit.edu.pem $HOMEWORLD_DIR/https-certs
193191
$ spire setup bootstrap-registry
194192

195193
## Confirm container launching
@@ -203,12 +201,13 @@ DNS, but when that happens, you can turn it back off:
203201

204202
Deploy flannel into the cluster:
205203

206-
$ cd deployment/deployment-config/cluster-gen/
207-
$ hyperkube kubectl create -f flannel.yaml
204+
$ mkdir cluster-gen
205+
$ spire config gen-kube cluster-gen
206+
$ spire kubectl create -f cluster-gen/flannel.yaml
208207

209208
Wait a bit for propagation.
210209

211-
$ hyperkube kubectl get pods --namespace=kube-system
210+
$ spire kubectl get pods --namespace=kube-system
212211
NAME READY STATUS RESTARTS AGE
213212
kube-flannel-ds-1r1cx 1/1 Running 0 49s
214213
kube-flannel-ds-2cxj5 1/1 Running 0 49s
@@ -219,7 +218,7 @@ Wait a bit for propagation.
219218

220219
Verify flannel functionality by running flannel tests on two different nodes:
221220

222-
$ # two nodes
221+
$ # on two different nodes
223222
$ ssh root@<worker>.mit.edu
224223
# rkt run --debug --interactive=true --net=rkt.kubernetes.io homeworld.mit.edu/debian
225224
$ ip addr # make sure this provides a 172.18 IP, and not a 172.16 IP.
@@ -231,11 +230,11 @@ If the ping works both ways, then flannel works! At least at a basic level.
231230

232231
Deploy dns-addon into the cluster:
233232

234-
$ hyperkube kubectl create -f dns-addon.yaml
233+
$ spire kubectl create -f dns-addon.yaml
235234

236235
Wait for deployment to succeed:
237236

238-
$ hyperkube kubectl get pods --namespace=kube-system
237+
$ spire kubectl get pods --namespace=kube-system
239238
NAME READY STATUS RESTARTS AGE
240239
kube-dns-v20-69lrg 3/3 Running 0 1m
241240
kube-dns-v20-clh2z 3/3 Running 0 1m

0 commit comments

Comments
 (0)