k8s - Volume 简介和HostPath的使用
K8S 的持久化
K8S 实现持久化存储的方法有很多种
例如 卷 (Volume), 持久卷(PV), 临时卷(EV) 等， 还有很多不常用的选项上图没有列出来
其中Volume 本身也分很多种
包括
Secret, configMap(之前的文章covered了)， hostPath, emptyDir等
本文主要focus on hostPath
HostPath 的简介
官方定义：
hostPath 卷能将主机节点文件系统上的文件或目录挂载到你的 Pod 中。 虽然这不是大多数 Pod 需要的，但是它为一些应用提供了强大的逃生舱。
简单来讲就是让k8s node 的目录or 文件map在你的POD 容器里
至于为什么上面提到的逃生舱， 是因为假如你的POD 崩溃了or 被shutdown， 我们仍然可以在node对应的path上找到我们想要的数据(前提是把相应的数据output 到hostpath）
但是 官方并不推荐使用hostpath 把数据输出到node上
建议用local PersisentVolume代替， 主要是安全风险的原因
原因：
如果你通过准入时的验证来限制对节点上特定目录的访问，这种限制只有在你额外要求所有 hostPath 卷的挂载都是只读的情况下才有效。如果你允许不受信任的 Pod 以读写方式挂载任意主机路径， 则该 Pod 中的容器可能会破坏可读写主机挂载卷的安全性。
无论 hostPath 卷是以只读还是读写方式挂载，使用时都需要小心，这是因为：
- 访问主机文件系统可能会暴露特权系统凭证（例如 kubelet 的凭证）或特权 API（例如容器运行时套接字）， 这些可以被用于容器逃逸或攻击集群的其他部分。
- 具有相同配置的 Pod（例如基于 PodTemplate 创建的 Pod）可能会由于节点上的文件不同而在不同节点上表现出不同的行为。
- hostPath 卷的用量不会被视为临时存储用量。 你需要自己监控磁盘使用情况，因为过多的 hostPath 磁盘使用量会导致节点上的磁盘压力。
HostPath 的类型
除了必需的 path 属性外，你还可以选择为 hostPath 卷指定 type。
| value | desc |
|---|---|
| “” | 空字符串（默认）用于向后兼容，这意味着在安装 hostPath 卷之前不会执行任何检查。 |
| DirectoryOrCreate | 如果在给定路径上什么都不存在，那么将根据需要创建空目录，权限设置为 0755，具有与 kubelet 相同的组和属主信息。 |
| Directory | 在给定路径上必须存在的目录。 |
| FileOrCreate | 如果在给定路径上什么都不存在，那么将在那里根据需要创建空文件，权限设置为 0644，具有与 kubelet 相同的组和所有权。 |
| File | 在给定路径上必须存在的文件。 |
| Socket | 在给定路径上必须存在的 UNIX 套接字。 |
| CharDevice | （仅 Linux Node） 在给定路径上必须存在的字符设备。 |
| BlockDevice | （仅 Linux Node） 在给定路径上必须存在的块设备。 |
例子
这个例子是测试多个pod同时往1个hostpath 写日志
添加filter 让其可以输出api 调用日志
我们先为springboot cloud-order service 增加1个filter
@Component
@WebFilter
@Order(1)
@Slf4j
public class InfoFilter implements Filter {
@Autowired
private String hostname;
@Override
public void init(FilterConfig filterConfig) throws ServletException {
// Initialization code
}
@Override
public void doFilter(ServletRequest request, ServletResponse response, FilterChain chain) throws IOException, ServletException {
HttpServletRequest httpRequest = (HttpServletRequest) request;
// Log the request information
log.info("API of hostname: {} has been called, Request Method: {}, Request URI: {}, Requested from : {}", hostname, httpRequest.getMethod(), httpRequest.getRequestURI(), httpRequest.getRemoteAddr());
// You can log more information as needed
// Call the next filter in the chain
chain.doFilter(request, response);
}
@Override
public void destroy() {
// Cleanup code
}
}
为k8s-node1 加上label
这个例子测试 多个POD 同时为1个hostpath 写日志
所以为了方便测试， 需要把所有PODs 都部署在同1个node上。
所以我们要为k8s-node1 加上label
gateman@MoreFine-S500:/app/logs$ kubectl label nodes k8s-node1 cloud-order=enabled
node/k8s-node1 labeled
gateman@MoreFine-S500:/app/logs$ kubectl get nodes --show-labels
NAME STATUS ROLES AGE VERSION LABELS
k8s-master Ready control-plane,master 190d v1.23.6 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,ingress=true,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-master,kubernetes.io/os=linux,node-role.kubernetes.io/control-plane=,node-role.kubernetes.io/master=,node.kubernetes.io/exclude-from-external-load-balancers=
k8s-node0 Ready <none> 190d v1.23.6 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,ingress=true,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node0,kubernetes.io/os=linux
k8s-node1 Ready <none> 190d v1.23.6 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,cloud-order=enabled,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node1,kubernetes.io/os=linux
k8s-node3 Ready <none> 169d v1.23.6 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node3,kubernetes.io/os=linux
可以见到， cloud-order=enabled 这个label 已经加上到了k8s-node1
在k8s-node1 提前创建1个folder
这个例子用的类型是Directory
当然如果用DirectoryOrCreate 是可以skip 这个步骤
gateman@MoreFine-S500:~/projects/coding/k8s-s/service-case/cloud-order$ gcloud compute ssh k8s-node1
WARNING: This command is using service account impersonation. All API calls will be executed as [[email protected]].
WARNING: This command is using service account impersonation. All API calls will be executed as [[email protected]].
No zone specified. Using zone [europe-west2-c] for instance: [k8s-node1].
External IP address was not found; defaulting to using IAP tunneling.
WARNING:
To increase the performance of the tunnel, consider installing NumPy. For instructions,
please see https://cloud.google.com/iap/docs/using-tcp-forwarding#increasing_the_tcp_upload_bandwidth
WARNING: This command is using service account impersonation. All API calls will be executed as [[email protected]].
Warning: Permanently added 'compute.7230880099421476498' (ED25519) to the list of known hosts.
Linux k8s-node1 5.10.0-32-cloud-amd64 #1 SMP Debian 5.10.223-1 (2024-08-10) x86_64
The programs included with the Debian GNU/Linux system are free software;
the exact distribution terms for each program are described in the
individual files in /usr/share/doc/*/copyright.
Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent
permitted by applicable law.
Last login: Sat Mar 9 19:39:42 2024 from 192.168.0.35
gateman@k8s-node1:~$ sudo su -
root@k8s-node1:~# mkdir /k8s-shared
root@k8s-node1:~# cd /k8s-shared/
root@k8s-node1:/k8s-shared# ls
root@k8s-node1:/k8s-shared# mkdir logs
root@k8s-node1:/k8s-shared# ls
logs
编写deployment yaml
deployment-cloud-order-hostpath.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
labels: # label of this deployment
app: cloud-order # custom defined
author: nvd11
name: deployment-cloud-order # name of this deployment
namespace: default
spec:
replicas: 3 # desired replica count, Please note that the replica Pods in a Deployment are typically distributed across multiple nodes.
revisionHistoryLimit: 10 # The number of old ReplicaSets to retain to allow rollback
selector: # label of the Pod that the Deployment is managing,, it's mandatory, without it , we will get this error
# error: error validating data: ValidationError(Deployment.spec.selector): missing required field "matchLabels" in io.k8s.apimachinery.pkg.apis.meta.v1.LabelSelector ..
matchLabels:
app: cloud-order
strategy: # Strategy of upodate
type: RollingUpdate # RollingUpdate or Recreate
rollingUpdate:
maxSurge: 25% # The maximum number of Pods that can be created over the desired number of Pods during the update
maxUnavailable: 25% # The maximum number of Pods that can be unavailable during the update
template: # Pod template
metadata:
labels:
app: cloud-order # label of the Pod that the Deployment is managing. must match the selector, otherwise, will get the error Invalid value: map[string]string{"app":"bq-api-xxx"}: `selector` does not match template `labels`
spec: # specification of the Pod
containers:
- image: europe-west2-docker.pkg.dev/jason-hsbc/my-docker-repo/cloud-order:1.1.0 # image of the container
imagePullPolicy: Always
name: container-cloud-order
command: ["bash"]
args:
- "-c"
- |
java -jar -Dserver.port=8080 app.jar --spring.profiles.active=$APP_ENVIRONMENT --logging.file.name=/app/logs/cloud-order.log
env: # set env varaibles
- name: APP_ENVIRONMENT # name of the environment variable
value: prod # value of the environment variable
volumeMounts:
- name: volume-log
mountPath: /app/logs/
readOnly: false # read only is set to false
ports:
- containerPort: 8080
name: cloud-order
nodeSelector:
cloud-order: enabled
volumes:
- name: volume-log
hostPath:
path: /k8s-shared/logs # the path on the host (k8s node)
type: Directory
restartPolicy: Always # Restart policy for all containers within the Pod
terminationGracePeriodSeconds: 10 # The period of time in seconds given to the Pod to terminate gracefully
注意几点，
- 启动命令加上 --logging.file.name=/app/logs/cloud-order.log 让日志输出到指定位置
- 加上volume-log 这个hostpath, 映射的主机路径是 /k8s-shared/logs
- volumemount 那里把 volume-log mount在了 /app/logs
- 使用nodeSelector 让pods 都部署在1个node
部署
gateman@MoreFine-S500:~/projects/coding/k8s-s/service-case/cloud-order/volumes$ kubectl delete deployment deployment-cloud-order
deployment.apps "deployment-cloud-order" deleted
gateman@MoreFine-S500:~/projects/coding/k8s-s/service-case/cloud-order/volumes$ kubectl apply -f deployment-cloud-order-hostpath.yaml
deployment.apps/deployment-cloud-order created
gateman@MoreFine-S500:~/projects/coding/k8s-s/service-case/cloud-order/volumes$ kubectl get pods
NAME READY STATUS RESTARTS AGE
deployment-bq-api-service-6f6ffc7866-8djx9 1/1 Running 3 (13d ago) 18d
deployment-bq-api-service-6f6ffc7866-g4854 1/1 Running 12 (13d ago) 61d
deployment-bq-api-service-6f6ffc7866-lwxt7 1/1 Running 14 (13d ago) 64d
deployment-bq-api-service-6f6ffc7866-mxwcq 1/1 Running 11 (13d ago) 61d
deployment-cloud-order-ff4989c97-h8ktj 1/1 Running 0 7s
deployment-cloud-order-ff4989c97-xg2x2 1/1 Running 0 7s
deployment-cloud-order-ff4989c97-zc2dq 1/1 Running 0 7s
测试
多次调用cloud-order的某个api
curl http://www.jp-gcp-vms.cloud:8085/cloud-order/actuator/info
curl http://www.jp-gcp-vms.cloud:8085/cloud-order/actuator/info
curl http://www.jp-gcp-vms.cloud:8085/cloud-order/actuator/info
curl http://www.jp-gcp-vms.cloud:8085/cloud-order/actuator/info
curl http://www.jp-gcp-vms.cloud:8085/cloud-order/actuator/info
进入k8s-node1 查看相应的日志文件：
root@k8s-node1:/k8s-shared/logs# tail -f cloud-order.log
...
2024-09-01T14:28:14.298Z INFO 1 --- [http-nio-8080-exec-1] com.home.clouduser.filter.InfoFilter : API of hostname: deployment-cloud-order-ff4989c97-h8ktj has been called, Request Method: GET, Request URI: /actuator/info, Requested from : 192.168.0.35
2024-09-01T14:28:15.384Z INFO 1 --- [http-nio-8080-exec-2] com.home.clouduser.filter.InfoFilter : API of hostname: deployment-cloud-order-ff4989c97-h8ktj has been called, Request Method: GET, Request URI: /actuator/info, Requested from : 192.168.0.35
2024-09-01T14:28:15.856Z INFO 1 --- [http-nio-8080-exec-2] com.home.clouduser.filter.InfoFilter : API of hostname: deployment-cloud-order-ff4989c97-zc2dq has been called, Request Method: GET, Request URI: /actuator/info, Requested from : 192.168.0.35
2024-09-01T14:28:17.211Z INFO 1 --- [http-nio-8080-exec-3] com.home.clouduser.filter.InfoFilter : API of hostname: deployment-cloud-order-ff4989c97-h8ktj has been called, Request Method: GET, Request URI: /actuator/info, Requested from : 192.168.0.35
2024-09-01T14:28:18.183Z INFO 1 --- [http-nio-8080-exec-4] com.home.clouduser.filter.InfoFilter : API of hostname: deployment-cloud-order-ff4989c97-h8ktj has been called, Request Method: GET, Request URI: /actuator/info, Requested from : 192.168.0.35
2024-09-01T14:28:19.128Z INFO 1 --- [http-nio-8080-exec-2] com.home.clouduser.filter.InfoFilter : API of hostname: deployment-cloud-order-ff4989c97-xg2x2 has been called, Request Method: GET, Request URI: /actuator/info, Requested from : 192.168.0.35
可以见到 ， 由于loadbalancer 的关系， 多个pods轮流被调用， 但是都可以写进同1个hostpath内的日志文件
所以如果用append 方式（例如写日子） 多个pod 写同1个host path 是不会由文件冲突问题的！