zookeeper和k8s_k8s部署有状态服务zookeeper示例
我们先不考虑配置文件的前提下:apiVersion: apps/v1kind: StatefulSet#####固定hostname,有状态的服务使用这个 statefalset有个问题,就是如果那个pod不是running状态,这个主机名是无法解析的,这样就构成了一个死循环,我sed替换主机名的时候由于pod还不是running状态,她只能获取自己的主机名。无法获取别人的主机名,所以在zo...
我们先不考虑配置文件的前提下:
apiVersion: apps/v1
kind: StatefulSet #####固定hostname,有状态的服务使用这个 statefalset有个问题,就是如果那个pod不是running状态,这个主机名是无法解析的,这样就构成了一个死循环,我sed替换主机名的时候由于pod还不是running状态,她只能获取自己的主机名。无法获取别人的主机名,所以在zookeeper中换成了换成了ip
metadata:
name: zookeeper
spec:
serviceName: zookeeper ####所以生成的3个pod的名字叫zookeeper-0,zookeeper-1,zookeeper-2
replicas: 3
revisionHistoryLimit: 10
selector: ##statefulset必须有的
matchLabels:
app: zookeeper
template:
metadata:
labels:
app: zookeeper
spec:
volumes:
- name: volume-logs
hostPath:
path: /var/log/zookeeper
containers:
- name: zookeeper
image: harbor.test.com/middleware/zookeeper:3.4.10
imagePullPolicy: IfNotPresent
livenessProbe:
tcpSocket:
port: 2181
initialDelaySeconds: 30
timeoutSeconds: 3
periodSeconds: 5
successThreshold: 1
failureThreshold: 2
ports:
- containerPort: 2181
protocol: TCP
- containerPort: 2888
protocol: TCP
- containerPort: 3888
protocol: TCP
env:
- name: SERVICE_NAME
value: "zookeeper"
- name: MY_POD_NAME #声明k8s自带的变量,这样在pod创建之后,在其中可以直接echo ${MY_POD_NAME}得到hostname
valueFrom:
fieldRef:
fieldPath: metadata.name
volumeMounts:
- name: volume-logs
mountPath: /var/log/zookeeper
nodeSelector:
zookeeper: enable
---
apiVersion: v1
kind: Service
metadata:
name: zookeeper #我的cluster名字为这个,在任意一个生成的pod中可以ping zookeeper,相当于zookeeper为生成的3个pod的cluster_name,会发现每次ping出的地址不一定相同,nslookup zookeeper得到的是3个pod的pod ip,共3条记录。
spec:
ports:
- port: 2181
selector:
app: zookeeper
clusterIP: None #此句必须加上
[root@host5 src]# kubectl get pod --all-namespaces -o wide
NAMESPACE NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
default zookeeper-0 1/1 Running 0 12m 192.168.55.69 host3
default zookeeper-1 1/1 Running 0 12m 192.168.31.93 host4
default zookeeper-2 1/1 Running 0 12m 192.168.55.70 host3
bash-4.3# nslookup zookeeper
nslookup: can't resolve '(null)': Name does not resolve
Name: zookeeper
Address 1: 192.168.55.70 zookeeper-2.zookeeper.default.svc.cluster.local
Address 2: 192.168.55.69 zookeeper-0.zookeeper.default.svc.cluster.local
Address 3: 192.168.31.93 zookeeper-1.zookeeper.default.svc.cluster.local
bash-4.3# ping zookeeper-0.zookeeper
PING zookeeper-0.zookeeper (192.168.55.69): 56 data bytes
64 bytes from 192.168.55.69: seq=0 ttl=63 time=0.109 ms
64 bytes from 192.168.55.69: seq=1 ttl=63 time=0.212 ms
^C
--- zookeeper-0.zookeeper ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 0.109/0.160/0.212 ms
bash-4.3# ping zookeeper-1.zookeeper
PING zookeeper-1.zookeeper (192.168.31.93): 56 data bytes
64 bytes from 192.168.31.93: seq=0 ttl=62 time=0.535 ms
64 bytes from 192.168.31.93: seq=1 ttl=62 time=0.507 ms
64 bytes from 192.168.31.93: seq=2 ttl=62 time=0.587 ms
^C
--- zookeeper-1.zookeeper ping statistics ---
3 packets transmitted, 3 packets received, 0% packet loss
round-trip min/avg/max = 0.507/0.543/0.587 ms
bash-4.3# ping zookeeper-2.zookeeper
PING zookeeper-2.zookeeper (192.168.55.70): 56 data bytes
64 bytes from 192.168.55.70: seq=0 ttl=64 time=0.058 ms
64 bytes from 192.168.55.70: seq=1 ttl=64 time=0.081 ms
^C
--- zookeeper-2.zookeeper ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 0.058/0.069/0.081 ms
k8s自带的常用变量如下:
env:
- name: MY_NODE_NAME
valueFrom:
fieldRef:
fieldPath: spec.nodeName
- name: MY_POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
- name: MY_POD_NAMESPACE
valueFrom:
fieldRef:
fieldPath: metadata.namespace
- name: MY_POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP
- name: MY_POD_SERVICE_ACCOUNT
valueFrom:
fieldRef:
fieldPath: spec.serviceAccountName
spec.nodeName : pod所在节点的IP、宿主主机IP
status.podIP :pod IP
我们再看配置文件:
[root@docker06 conf]# cat zoo.cfg |grep -v ^#|grep -v ^$
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/data
clientPort=2181
clientPortAddress= docker06
server.1=docker05:2888:3888
server.2=docker06:2888:3888
server.3=docker04:2888:3888
snapCount=10000
leaderServes=yes
autopurge.snapRetainCount=3
autopurge.purgeInterval=2
maxClientCnxns=1000
我们需要修改成形如:
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/data
clientPort=2181
clientPortAddress= docker06 #下面的3行是固定的,主要是这行需要修改成本机的MY_POD_IP,我们可以用configmap挂载配置文件,然后在pod里面用sed替换掉这行配置
server.1=zookeeper-0.zookeeper:2888:3888
server.2=zookeeper-1.zookeeper:2888:3888
server.3=zookeeper-2.zookeeper:2888:3888
snapCount=10000
leaderServes=yes
autopurge.snapRetainCount=3
autopurge.purgeInterval=2
maxClientCnxns=1000
参考如下这种方式:
先将配置文件通过configmap挂载进pod里面,如fix-ip.sh
apiVersion: v1
kind: ConfigMap
metadata:
name: redis-cluster
labels:
app: redis-cluster
data:
fix-ip.sh: |
#!/bin/sh
CLUSTER_CONFIG="/var/lib/redis/nodes.conf"
if [ -f ${CLUSTER_CONFIG} ]; then
if [ -z "${POD_IP}" ]; then
echo "Unable to determine Pod IP address!"
exit 1
fi
echo "Updating my IP to ${POD_IP} in ${CLUSTER_CONFIG}"
sed -i.bak -e "/myself/s/[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}/${POD_IP}/" ${CLUSTER_CONFIG}
fi
exec "$@"
redis.conf: |+
cluster-enabled yes
cluster-require-full-coverage no
cluster-node-timeout 15000
cluster-config-file /var/lib/redis/nodes.conf
cluster-migration-barrier 1
appendonly yes
protected-mode no
然后在启动pod的时候执行这个脚本:
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: redis-cluster
labels:
app: redis-cluster
spec:
serviceName: redis-cluster
replicas: 6
selector:
matchLabels:
app: redis-cluster
template:
metadata:
labels:
app: redis-cluster
spec:
containers:
- name: redis
image: 10.11.100.85/library/redis
ports:
- containerPort: 6379
name: client
- containerPort: 16379
name: gossip
command: ["/etc/redis/fix-ip.sh", "redis-server", "/etc/redis/redis.conf"] #此处先执行了那个脚本,然后启动的redis
readinessProbe:
exec:
command:
- sh
- -c
- "redis-cli -h $(hostname) ping"
initialDelaySeconds: 15
timeoutSeconds: 5
livenessProbe:
exec:
command:
- sh
- -c
- "redis-cli -h $(hostname) ping"
initialDelaySeconds: 20
periodSeconds: 3
env:
- name: POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP
volumeMounts:
- name: conf
mountPath: /etc/redis
readOnly: false
- name: data
mountPath: /var/lib/redis
readOnly: false
volumes:
- name: conf
configMap:
name: redis-cluster
defaultMode: 0755
# items:
# - key: redis.conf
# path: redis.conf
# - key: fix-ip.sh
# path: fix-ip.sh
volumeClaimTemplates:
- metadata:
name: data
labels:
name: redis-cluster
spec:
accessModes: [ "ReadWriteOnce" ]
resources:
requests:
storage: 150Mi
注意:通过configmap生成的配置文件为只读,无法通过sed修改,可以通过挂载到临时目录,然后拷过去之后sed,但是这样也存在一个问题,就是你动态修改了configmap,只会改变临时目录里的文件,而不会改变考过去的文件
实际生产环境的配置:
1.重新制作image
[root@host4 zookeeper]# ll
总用量 4
drwxr-xr-x 2 root root 45 5月 24 15:48 conf
-rw-r--r-- 1 root root 143 5月 23 06:19 Dockerfile
drwxr-xr-x 2 root root 20 5月 24 15:48 scripts
[root@host4 conf]# cd conf
[root@host4 conf]# ll
总用量 8
-rw-r--r-- 1 root root 1503 5月 23 04:15 log4j.properties
-rw-r--r-- 1 root root 324 5月 24 15:48 zoo.cfg
[root@host4 conf]# cat zoo.cfg
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/data
clientPort=2181
clientPortAddress=PODIP #此处用ip,下面用主机名,原因见本文上面
server.1=zookeeper-0.zookeeper:2888:3888
server.2=zookeeper-1.zookeeper:2888:3888
server.3=zookeeper-2.zookeeper:2888:3888
snapCount=10000
leaderServes=yes
autopurge.snapRetainCount=3
autopurge.purgeInterval=2
maxClientCnxns=1000
[root@host4 conf]# cd ../scripts/
[root@host4 scripts]# ll
总用量 4
-rwxr-xr-x 1 root root 177 5月 24 15:48 sed.sh
[root@host4 scripts]# cat sed.sh
#!/bin/bash
MY_ID=`echo ${MY_POD_NAME} |awk -F'-' '{print $NF}'`
MY_ID=`expr ${MY_ID} + 1`
echo ${MY_ID} > /data/myid
sed -i 's/PODIP/'${MY_POD_IP}'/g' /conf/zoo.cfg
exec "$@"
[root@host4 scripts]# cd ..
[root@host4 zookeeper]# ls
conf Dockerfile scripts
[root@host4 zookeeper]# cat Dockerfile
FROM harbor.test.com/middleware/zookeeper:3.4.10
MAINTAINER rongruixue@163.com
ARG zookeeper_version=3.4.10
COPY conf /conf/
COPY scripts /
这样我们docker build就制作出了image :harbor.test.com/middleware/zookeeper:v3.4.10
然后我们启通过yml启动pod:
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: zookeeper
spec:
# podManagementPolicy: Parallel #此配置决定是否让3个pod同时起来,而不是按 0 1 2的顺序
serviceName: zookeeper
replicas: 3
revisionHistoryLimit: 10
selector:
matchLabels:
app: zookeeper
template:
metadata:
labels:
app: zookeeper
spec:
volumes:
- name: volume-logs
hostPath:
path: /var/log/zookeeper
- name: volume-data
hostPath:
path: /opt/zookeeper/data
terminationGracePeriodSeconds: 10
containers:
- name: zookeeper
image: harbor.test.com/middleware/zookeeper:v3.4.10
imagePullPolicy: Always
ports:
- containerPort: 2181
protocol: TCP
- containerPort: 2888
protocol: TCP
- containerPort: 3888
protocol: TCP
env:
- name: SERVICE_NAME
value: "zookeeper"
- name: MY_POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP
- name: MY_POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
volumeMounts:
- name: volume-logs
mountPath: /var/log/zookeeper
#- name: volume-data 此处不能挂载、data到本地,否则如果两个pod分配到同一个节点会相互覆盖,myid也会被覆盖
# mountPath: /data
command:
- /bin/bash
- -c
- -x
- |
/sed.sh #此脚本作用就是讲podip写入zoo.cfg配置文件中,然后写/data/myid
sleep 10
zkServer.sh start-foreground
nodeSelector:
zookeeper: enable
---
apiVersion: v1
kind: Service
metadata:
name: zookeeper
spec:
ports:
- port: 2181
selector:
app: zookeeper
clusterIP: None
更多推荐
所有评论(0)