Skip to content

Commit 4b3fc62

Browse files
mumutyalroot
mumutyal
and
root
authored
Portworx utility scripts (#103)
* Adding px util scripts * Adding px utils * Adding portworx utility scripts * Readme change * PX utility script Co-authored-by: root <[email protected]>
1 parent 117ef87 commit 4b3fc62

File tree

9 files changed

+993
-0
lines changed

9 files changed

+993
-0
lines changed

‎px-utils/README.md‎

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,8 @@
1+
# px-utility-scripts
2+
This is collection of Portworx utilities
3+
4+
5+
6+
7+
8+

‎px-utils/px_cleanup/README.md‎

Lines changed: 31 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,31 @@
1+
# px-cleanup
2+
3+
The PX Cleanup script works in the following way.
4+
5+
When uninstalling Portworx from cluster, There are 2 options:
6+
7+
Stop Portworx and remove the Kubernetes specs and completely wipe the data.
8+
you will be prompted for confirmation if yes then removes the volumes data also
9+
10+
2. Stop Portworx and remove the Kubernetes specs without wiping the data.
11+
12+
Uninstalling or deleting the Portworx daemonset only removes the Portworx containers from the nodes. As the configurations files which Portworx use are persisted on the nodes the storage devices and the data volumes are still intact. These Portworx volumes can be used again if the Portworx containers are started with the same configuration files. (edited)
13+
14+
15+
Prerequisites
16+
17+
1) Set the cluster config using following command
18+
19+
export KUBECONFIG="Path to the config yaml file"
20+
21+
22+
Once the cluster config is set run the scirpt
23+
24+
25+
sudo ./px_cleanup.sh
26+
27+
28+
29+
30+
31+

‎px-utils/px_cleanup/px_cleanup.sh‎

Lines changed: 267 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,267 @@
1+
#!/bin/bash
2+
3+
TALISMAN_IMAGE=portworx/talisman
4+
TALISMAN_TAG=latest
5+
WIPE_CLUSTER="--wipecluster"
6+
MAX_RETRIES=60
7+
TIME_BEFORE_RETRY=5 #seconds
8+
JOB_NAME=talisman
9+
KUBECTL_EXTRA_OPTS=""
10+
WIPER_IMAGE=portworx/px-node-wiper
11+
WIPER_TAG=latest
12+
13+
14+
logmessage() {
15+
echo "" 2>&1
16+
echo "$@" 2>&1
17+
}
18+
19+
if ! which jq &>/dev/null
20+
then
21+
echo "Jq is not installed... exiting"
22+
exit 1
23+
fi
24+
25+
##Check ibmcloud instaled or not
26+
if ! which ibmcloud &>/dev/null
27+
then
28+
echo "IBM Cloud is not installed. Please install ibmcloud..... exiting"
29+
exit 1
30+
fi
31+
32+
33+
ask() {
34+
# https://djm.me/ask
35+
local prompt default reply
36+
prompt="Y/n"
37+
default=N
38+
39+
# Ask the question (not using "read -p" as it uses stderr not stdout)<Paste>
40+
echo -n "$1 [$prompt]:"
41+
42+
# Read the answer (use /dev/tty in case stdin is redirected from somewhere else)
43+
read reply </dev/tty
44+
if [ $? -ne 0 ]; then
45+
logmessage "ERROR: Could not ask for user input - please run via interactive shell"
46+
fi
47+
48+
# Default? (e.g user presses enter)
49+
if [ -z "$reply" ]; then
50+
reply=$default
51+
fi
52+
53+
# Check if the reply is valid
54+
case "$reply" in
55+
Y*|y*) return 0 ;;
56+
N*|n*) return 1 ;;
57+
* ) echo "invalid reply: $reply"; return 1 ;;
58+
esac
59+
}
60+
CLUSTER_NAME=kubectl -n kube-system get cm cluster-info -o jsonpath='{.data.cluster-config\.json}' | jq -r '.name'
61+
62+
if ! ask "The operation will delete Portworx components and metadata from the cluster. Do you want to continue?" N; then
63+
logmessage "Aborting Portworx wipe from the cluster..."
64+
exit 1
65+
else
66+
if ! ask "Do you want to wipeout the data also from the volumes . Please enter?" N; then
67+
logmessage "The operation will delete Portworx components and metadata from the cluster.The data will not be wiped out fromm the voluems..."
68+
WIPE_CLUSTER=""
69+
else
70+
if ! ask "The operation will delete Portworx components and metadata from the cluster. The operation is irreversible and will lead to DATA LOSS. Do you want to continue?" N; then
71+
logmessage "The operation will delete Portworx components and metadata from the cluster.The data will not be wiped out fromm the voluems..."
72+
WIPE_CLUSTER=""
73+
else
74+
logmessage "The operation will delete Portworx components and metadata and the data on the volumes..."
75+
WIPE_CLUSTER="--wipecluster"
76+
fi
77+
fi
78+
fi
79+
80+
##########################
81+
# Added IS_OPENSHIFT variable to check if cluster is an openshift cluster
82+
##########################
83+
if [[ $CLUSTER_VERSION == *"openshift"* ]]; then
84+
command -v oc
85+
if [ $? -eq 0 ]; then
86+
_out=$(command -v oc 2>&1 && oc version 2>&1)
87+
if [ $? -eq 0 ]; then
88+
echo "Detected OpenShift system. Adding talisman-account user to privileged scc"
89+
oc adm policy add-scc-to-user privileged system:serviceaccount:kube-system:talisman-account
90+
if [ $? -ne 0 ]; then
91+
logmessage "failed to add talisman-account to privileged scc. exit code: $?"
92+
fi
93+
else
94+
echo "oc binary found but oc version command failed. Not using OpenShift mode."
95+
fi
96+
fi
97+
fi
98+
##########################
99+
100+
VER=$(kubectl version --short | awk -Fv '/Server Version: /{print $3}')
101+
if [ -z "$VER" ]; then
102+
logmessage "failed to get kubernetes server version. Make sure you have kubectl setup on current machine."
103+
fi
104+
105+
106+
VER=( ${VER//./ } )
107+
echo "Parsed version is "${VER[0]}.${VER[1]}""
108+
109+
if [ "${VER[0]}.${VER[1]}" == "1.7" ] || [ "${VER[0]}.${VER[1]}" == "1.6" ]; then
110+
logmessage "This script doesn't support wiping Portworx from Kubernetes $VER clusters."
111+
fi
112+
113+
CLIENT_VER=$(kubectl version --short | awk -Fv '/Client Version: /{print $2}')
114+
if [ -z "$CLIENT_VER" ]; then
115+
logmessage "failed to get kubernetes client version. Make sure you have kubectl setup on current machine."
116+
fi
117+
118+
CLIENT_VERI=$(echo $CLIENT_VER | awk -F. '{print $1*100+$2}')
119+
echo "Parsed client version (int) is ${CLIENT_VERI}"
120+
if [ $CLIENT_VERI -lt 114 ]; then
121+
KUBECTL_EXTRA_OPTS="--show-all"
122+
else
123+
KUBECTL_EXTRA_OPTS=""
124+
fi
125+
126+
kubectl delete -n kube-system job talisman 2>/dev/null || true
127+
128+
RETRY_CNT=0
129+
while true; do
130+
PODS=$(kubectl get pods -n kube-system -l name=$JOB_NAME $KUBECTL_EXTRA_OPTS 2>/dev/null)
131+
if [ $? -eq 0 ]; then
132+
NUM_PODS=$(echo -n "$PODS" | grep -c -v NAME)
133+
if [ $NUM_PODS -eq 0 ]; then
134+
break
135+
fi
136+
fi
137+
138+
RETRY_CNT=$((RETRY_CNT+1))
139+
if [ $RETRY_CNT -ge $MAX_RETRIES ]; then
140+
logmessage "failed to delete old talisman pods"
141+
fi
142+
143+
sleep $TIME_BEFORE_RETRY
144+
done
145+
146+
147+
cat <<EOF | kubectl apply -f -
148+
---
149+
apiVersion: v1
150+
kind: ServiceAccount
151+
metadata:
152+
name: talisman-account
153+
namespace: kube-system
154+
---
155+
kind: ClusterRoleBinding
156+
apiVersion: rbac.authorization.k8s.io/v1
157+
metadata:
158+
name: talisman-role-binding
159+
subjects:
160+
- kind: ServiceAccount
161+
name: talisman-account
162+
namespace: kube-system
163+
roleRef:
164+
kind: ClusterRole
165+
name: cluster-admin
166+
apiGroup: rbac.authorization.k8s.io
167+
---
168+
169+
apiVersion: batch/v1
170+
kind: Job
171+
metadata:
172+
name: $JOB_NAME
173+
namespace: kube-system
174+
spec:
175+
backoffLimit: 1
176+
template:
177+
metadata:
178+
labels:
179+
name: $JOB_NAME
180+
spec:
181+
serviceAccount: talisman-account
182+
containers:
183+
- name: $JOB_NAME
184+
image: $TALISMAN_IMAGE:$TALISMAN_TAG
185+
args: ["-operation",
186+
"delete",
187+
"$WIPE_CLUSTER",
188+
"-wiperimage",
189+
"$WIPER_IMAGE",
190+
"-wipertag",
191+
"$WIPER_TAG"]
192+
imagePullPolicy: Always
193+
volumeMounts:
194+
- name: etcpwx
195+
mountPath: /etc/pwx
196+
volumes:
197+
- name: etcpwx
198+
hostPath:
199+
path: /etc/pwx
200+
restartPolicy: Never
201+
EOF
202+
203+
echo "Talisman job for wiping Portworx started. Monitor logs using: 'kubectl logs -n kube-system -l job-name=talisman'"
204+
205+
NUM_DESIRED=1
206+
RETRY_CNT=0
207+
while true; do
208+
NUM_SUCCEEDED=0
209+
NUM_FAILED=0
210+
PODS=$(kubectl get pods -n kube-system -l name=$JOB_NAME 2>/dev/null)
211+
if [ $? -eq 0 ]; then
212+
CREATING=$(echo "$PODS" | grep ContainerCreating)
213+
if [ ! -z "$CREATING" ]; then
214+
echo "Pod that will perform wipe of Portworx is still in container creating phase"
215+
else
216+
NUM_FAILED=$(kubectl get job -n kube-system talisman $KUBECTL_EXTRA_OPTS -o jsonpath='{.status.failed}' 2>/dev/null)
217+
if [ $? -eq 0 ]; then
218+
if [ ! -z "$NUM_FAILED" ] && [ $NUM_FAILED -ge 1 ]; then
219+
kubectl logs -n kube-system -l name=$JOB_NAME
220+
logmessage "Job to wipe px cluster failed."
221+
fi
222+
fi
223+
224+
NUM_SUCCEEDED=$(kubectl get job -n kube-system talisman $KUBECTL_EXTRA_OPTS -o jsonpath='{.status.succeeded}' 2>/dev/null)
225+
if [ ! -z "$NUM_SUCCEEDED" ] && [ $NUM_SUCCEEDED -eq $NUM_DESIRED ]; then
226+
break
227+
fi
228+
229+
echo "waiting on $JOB_NAME to complete..."
230+
RUNNING_POD=$(echo "$PODS" | grep Running | awk '/^talisman/{print $1}')
231+
if [ ! -z "$RUNNING_POD" ]; then
232+
echo "Monitoring logs of pod: $RUNNING_POD"
233+
kubectl logs -n kube-system -f $RUNNING_POD
234+
fi
235+
fi
236+
fi
237+
238+
RETRY_CNT=$((RETRY_CNT+1))
239+
if [ $RETRY_CNT -ge $MAX_RETRIES ]; then
240+
kubectl logs -n kube-system -l name=$JOB_NAME
241+
logmessage "Timed out trying to wipe Portworx cluster."
242+
fi
243+
244+
sleep $TIME_BEFORE_RETRY
245+
done
246+
247+
echo "Portworx cluster wipe succesfully completed."
248+
_rc=0
249+
kubectl delete job -n kube-system talisman || _rc=$?
250+
kubectl delete serviceaccount -n kube-system talisman-account || _rc=$?
251+
kubectl delete clusterrolebinding talisman-role-binding || _rc=$?
252+
kubectl delete crd volumeplacementstrategies.portworx.io
253+
if [ $_rc -ne 0 ]; then
254+
logmessage "error cleaning up pods"
255+
fi
256+
echo "removing the portworx helm from the cluster"
257+
_rc=0
258+
helm_release=$(helm ls -A --output json | jq -r '.[]|select(.name=="portworx") | .name')
259+
[[ -z "$helm_release" ]] && { echo "Unable to find helm release for portworx. Ensure your helm client is at version 3 and has access to the cluster."; exit; }
260+
helm uninstall portworx || _rc=$?
261+
if [ $_rc -ne 0 ]; then
262+
logmessage "error removing the helm relese"
263+
exit 1;
264+
fi
265+
echo "Removing the Service from the catalog"
266+
Bx_PX_svc_name=$(ibmcloud resource service-instances --service-name portworx --output json | jq -r --arg CLUSTERNAME $CLUSTER_NAME '.[]|select((.parameters.clusters==$CLUSTERNAME)) | .name')
267+
ibmcloud resource service-instance-delete "${Bx_PX_svc_name}" -f

‎px-utils/px_install/README.md‎

Lines changed: 22 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,22 @@
1+
# px-install
2+
3+
The PX Intall script intall the portworx service in the cluster and create the service in catalog
4+
5+
6+
Prerequisites
7+
8+
1) Set the cluster config using following command
9+
10+
export KUBECONFIG="Path to the config yaml file"
11+
12+
13+
Once the cluster config is set run the scirpt
14+
15+
16+
sudo ./px_install.sh clustername
17+
18+
19+
20+
21+
22+

0 commit comments

Comments
 (0)