This post mainly records the script for creating a community-edition JuiceFS PVC, to make configuration quick. For component deployment, see Managing Data in Kubernetes with Fluid and JuiceFS.
1. Set Environment Variables
1
2
3
4
5
6
| export ACCESS_KEY=
export SECRET_KEY=
export BUCKET=
export ENDPOINT=ks3-cn-beijing-internal.ksyun.com
export BUCKET_ENPOINT=$BUCKET.$ENDPOINT
export PROVIDER=ks3
|
1
2
3
4
5
| export NAMESPACE=
export PVC_NAME=
export NODE_SELECTOR_KEY=
export NODE_SELECTOR_VALUE=
|
export JUICEFS_IMAGE=juicedata/juicefs-fuse
export DEMO_IMAGE=shaowenchen/demo:ubuntu
For Redis configuration
1
2
3
4
| export REDIS_PASSWORD=
#ip:port/database
export REDIS_ENDPOINT=
export META_SERVER=redis://:$REDIS_PASSWORD@$REDIS_ENDPOINT
|
For PostgreSQL configuration
1
2
3
4
5
6
| export POSTGRES_USER=admin
export POSTGRES_PASSWORD=
export POSTGRES_IP=
export POSTGRES_PORT=5432
export POSTGRES_DB=
export META_SERVER=postgres://$POSTGRES_USER:$POSTGRES_PASSWORD@$POSTGRES_IP:$POSTGRES_PORT/$POSTGRES_DB?sslmode=disable
|
1
| psql -U $POSTGRES_USER -h $POSTGRES_IP -p $POSTGRES_PORT -d postgres -c "CREATE DATABASE $POSTGRES_DB;"
|
2. Initialize JuiceFS
1
2
3
4
5
6
| juicefs format \
--trash-days=1 \
--storage $PROVIDER \
--bucket $BUCKET_ENPOINT \
$META_SERVER \
$BUCKET
|
A trash-days of 0 disables the trash bin.
3. Create the Secret
1
2
3
4
5
6
7
8
9
10
11
12
| kubectl apply -f - <<EOF
apiVersion: v1
kind: Secret
metadata:
name: $PVC_NAME
namespace: $NAMESPACE
type: Opaque
stringData:
metaurl: $META_SERVER
access-key: $ACCESS_KEY
secret-key: $SECRET_KEY
EOF
|
4. Create the Dataset
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
| kubectl apply -f - <<EOF
apiVersion: data.fluid.io/v1alpha1
kind: Dataset
metadata:
name: $PVC_NAME
namespace: $NAMESPACE
spec:
accessModes:
- ReadWriteMany
mounts:
- name: $BUCKET
mountPoint: "juicefs:///"
options:
bucket: $BUCKET_ENPOINT
storage: $PROVIDER
encryptOptions:
- name: metaurl
valueFrom:
secretKeyRef:
name: $PVC_NAME
key: metaurl
- name: access-key
valueFrom:
secretKeyRef:
name: $PVC_NAME
key: access-key
- name: secret-key
valueFrom:
secretKeyRef:
name: $PVC_NAME
key: secret-key
EOF
|
5. Create the JuiceFSRuntime
Remember to adjust nodeSelector, because the worker consumes a large amount of resources.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
| kubectl apply -f - <<EOF
apiVersion: data.fluid.io/v1alpha1
kind: JuiceFSRuntime
metadata:
name: $PVC_NAME
namespace: $NAMESPACE
spec:
replicas: 1
juicefsVersion:
image: $JUICEFS_IMAGE
imageTag: ce-v1.1.0
fuse:
image: $JUICEFS_IMAGE
imageTag: ce-v1.1.0
cleanPolicy: OnDemand
worker:
nodeSelector:
$NODE_SELECTOR_KEY: "$NODE_SELECTOR_VALUE"
resources:
limits:
cpu: 15
memory: 100Gi
requests:
cpu: 1
memory: 10Gi
tieredstore:
levels:
- mediumtype: SSD
path: /data/jfs/cache
quota: 40960 # 40GiB
EOF
|
For the enterprise edition of JuiceFS, cache-group can be configured in the fuse and worker sections.
1
2
| options:
"cache-group": "$PVC_NAME"
|
Note that only nodes under the same high-speed network can use the same cache-group.
6. Create a Test Pod
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
| kubectl apply -f - <<EOF
apiVersion: apps/v1
kind: Deployment
metadata:
name: ${PVC_NAME}-deploy
namespace: $NAMESPACE
spec:
replicas: 1
selector:
matchLabels:
app: ${PVC_NAME}-demo
template:
metadata:
labels:
app: ${PVC_NAME}-demo
spec:
containers:
- name: demo
image: $DEMO_IMAGE
volumeMounts:
- mountPath: /data/jfs
name: data
nodeSelector:
$NODE_SELECTOR_KEY: "$NODE_SELECTOR_VALUE"
volumes:
- name: data
persistentVolumeClaim:
claimName: $PVC_NAME
EOF
|
1
| kubectl -n $NAMESPACE exec -it deploy/$PVC_NAME-deploy -- bash
|
1
| curl -sSL https://d.juicefs.com/install | sh -
|
