external etcd backup & restore

external etcd를 backup하고, restore한다.

 

작업순서

1. busybox pod을 생성

2. etcd backup

3. pod delete

4. restore

 

 

1. busybox pod을 생성한다.

yaml file 작성

apiVersion: v1
kind: Pod
metadata:
  name: busybox
  namespace: default
spec:
  containers:
  - image: busybox
    command:
      - sleep
      - "3600"
    imagePullPolicy: IfNotPresent
    name: busybox
  restartPolicy: Always

 

pod 생성

[root@km ~]# kubectl create -f busy.yaml
pod/busybox created

[root@km ~]# kubectl get pod
NAME      READY   STATUS    RESTARTS   AGE
busybox   1/1     Running   0          7s

 

2. etcd backup 진행

tmp 디렉토리에 etcd-backup.db 파일이 생성된다.

ETCDCTL_API=3 etcdctl snapshot save /tmp/etcd-backup.db --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key --endpoints=172.30.1.146:2379

 

3.pod delete

busybox pod을 삭제한다.

[root@km ~]# kubectl delete po busybox
pod "busybox" deleted

 

4. restore

4.1 snapshot을 사용해 /var/lib/etcd-backup 경로에 푼다.

[root@etcd etcd]# ETCDCTL_API=3 etcdctl snapshot restore /tmp/etcd-backup.db --data-dir /var/lib/etcd-backup
2022-09-17 23:40:29.939520 I | mvcc: restore compact to 18798
2022-09-17 23:40:29.946602 I | etcdserver/membership: added member 8e9e05c52164694d [http://localhost:2380] to cluster cdf818194e3a8c32

 

4.2 etcd 중지

etcd가 중지되면 k8s cluster에서 pod, svc 등의 정보를 조회할 수 없게 된다.

kubectl get pod 명령을 수행하면 hang 상태

systemctl stop etcd

 

4.3 기존 data-dir 삭제

rm -rf /var/lib/etcd/*

 

4.4 backup data-dir의 내용을 data-dir로 복사

cp /var/lib/etcd-backup/* /var/lib/etcd/

 

4.5 etcd 시작

systemctl start etcd

 

5. 복구 확인

확인해보니 삭제한 pod이 복구되었다.

[root@km pods]# kubectl get pod
NAME      READY   STATUS    RESTARTS   AGE
busybox   1/1     Running   0          3h59m