How to run ZenTao in Kubernetes with NFS

1. Test in Virtual host

Files needed to be edit

/opt/zbox/app/zentao/config/my.php
/opt/zbox/app/zentaoep/config/my.php
/opt/zbox/app/zentaopro/config/my.php
/opt/zbox/etc/php/php.ini
# Edit Port in it
/opt/zbox/etc/apache/httpd.conf

Create Group and User

Otherwise there might be AH00544: httpd: bad group name nogroup

useradd nobody
groupadd nogroup

Keep or rm them on your own

# Where attachments are
/opt/zbox/app/zentao/www/data/upload
# ZenTao auto backup 
root/opt/zbox/app/zentao/tmp/backup

Edit the above file and replace database as yours, I am using Alicloud RDS here.

After editting, start MySQL database service

/opt/zbox/zbox start or,

sh run/apache/apachectl start or /opt/zbox/run/apache/httpd -k start

2. Dependencies work before run container

Since I have uploaded some pictures before, I need to use storage here.

Create a script in /opt/run and link storage file to /opt/zbox

Install nfs

# Install nfs-utils
yum install-y nfs-utils
# Edit exports and grant privilege
cat>>/etc/exports<<EOF 
/opt/ 192.168.0.0/16(rw,no_root_squash,sync) 
EOF
# Start nfs service
systemctlenablerpcbind.service 
systemctlenablenfs-server.service 
systemctl start rpcbind.service 
systemctl start nfs-server.service 
systemctl stop firewalld 
systemctl disable firewalld 
rpcinfo -p
# exportfs nfs
exportfs -r 
exportfs

Create dockerfile

Since I want to link /opt/zbox/run with /opt/

So I need to keep one service in background

Create start.sh in /opt/run/

nohup/opt/run/apache/httpd -k start&
tail-999f /opt/run/nohup.out

Create dockerfile in /opt/run/

FROM harbor.aircourses.com/kubernetes/jdk-1.8.0_161:v1.1 
MAINTAINER PDABC Enterprise Container Images <[email protected]> 
RUN (useradd nobody;groupadd nogroup ) 
COPY run /opt/run 
ENV LANG=en_US.UTF-8 
ENV LANGUAGE=en_US.UTF-8 
ENTRYPOINT ["sh", "/opt/run/start.sh"]

Docker build and push

docker build -f Dockerfile-zbox-80 -t harbor.aircourses.com/kubernetes/zbox:v3
docker push harbor.aircourses.com/kubernetes/zbox:v3

3. Test in Kubernetes

Create pv.yaml

# Create nfs-PV
---
apiVersion:v1
kind:PersistentVolume
metadata:
     name:nfs-pv
     namespace:default
     labels:
        pv:nfs-pv
spec:
     capacity:
        storage:100Gi
     accessModes:
        -  ReadWriteManypersistentVolumeReclaimPolicy:Retain
     nfs:
        path:/opt/zbox
        server:192.168.13.212

kubectl create -f pv.yaml

Create pvc.yaml

# Create NFS-pvc
---
kind:PersistentVolumeClaim
apiVersion:v1
metadata:
     name:nfs-pvc
     namespace:default
spec:
     accessModes:
        -  ReadWriteMany
        resources:
        requests:
            storage:100Gi
     selector:
        matchLabels:
            pv:nfs-pv

kubectl apply -f pvc.yaml

Create deployment.yaml

apiVersion:apps/v1
kind:Deployment
metadata:
# deployment name has nothing to do with svc and ingress binding
    name:  zbox-dp
    namespace:default
spec:
    replicas:1
    # labels is necessary when create template since Deployment.spec.selector is required field and it has to correspond template.labels
    selector:
        matchLabels:
             app:zbox
    # template defined details will apply to all the copies ( e.g. deployment pod ), labels can't be defined in template.spec.containers.Use "kubectl get pods --show-labels" to check
    template:
        metadata:
             labels:
                app:zbox
        spec:
             containers:
# It doesn't matter about containers name and svc binding with ingress
             -  name:zbox
                image:harbor.aircourses.com/kubernetes/zbox:v4
                volumeMounts:
                -  mountPath:/opt/zbox
                   name:zbox-data
                ports:
                -  name:HTTP
                   containerPort:80
             volumes:
             -  name:zbox-data
                 persistentVolumeClaim:
                     claimName:nfs-pvc
             imagePullSecrets:
                 -  name:myregistrykey

kubectl apply -f deployment.yaml

Create svc-zbox.yaml

apiVersion:v1
kind:  Servicekubectl apply -f pvc.yaml
metadata:
    name:zbox
    namespace:default
spec:
    type:ClusterIP
    selector:
        app:zbox
    ports:
    -  name:HTTP
       port:80
       targetPort:80

kubectl apply -f svc-zbox.yaml

Create ingress-zbox.yaml

apiVersion:apps/v1
kind:  Deployment
metadata:
#deployment name has nothing to do with svc and ingress binding
    name:zbox-dp
    namespace:default
spec:
    replicas:1
# labels are necessary when creating template since Deployment.spec.selector is a required field and it has to correspond template. labels
    selector:
        matchLabels:
             app:zbox
# template defined details will apply to all the copies ( e.g. deployment pod ), labels can't be defined in template.spec.containers.Use "kubectl get pods --show-labels" to check
    template:
        metadata:
             labels:
                app:zbox
        spec:
             containers:
# It doesn't matter about containers name and svc binding with ingress
             -  name:zbox
                image:harbor.aircourses.com/jiaminxu/zbox:v1.0
                volumeMounts:
                -  mountPath:  /opt/zbox
                   name:zbox-data
                ports:
                -  name:HTTP
                   containerPort:80
             volumes:
             -  name:zbox-data
                persistentVolumeClaim:
                    claimName:nfs-pvc
             imagePullSecrets:
                -  name:myregistrykey