使用velero管理Kubernetes资源
使用velero管理Kubernetes资源
背景
经常在开发测试环境碰到这样一种情况，yaml文件被修改了，或者service被人删除了。这个时候需要第一时间恢复业务，保障其开发测试的可用性，在生产环境中，备份显得尤为重要，任何一个误操作都有可能导致业务受到影响。
一、简介
是 VMWare 开源的 k8s 集群备份、迁移工具。可以帮助我们完成 k8s 的例行备份工作，以便在出现上面问题的时候可以快速进行恢复。同时也提供了集群迁移功能，可以将 k8s 资源迁移到其他 k8s 集群的功能。Velero 将集群资源保存在对象存储中，默认情况下可以使用 AWS、Azure、GCP 的对象存储。这里使用aws的插件实现腾讯云cos存储（兼容s3）。
特性
- 集群备份
- 备份调度
- 备份钩子
二、前置条件
-
安装好
kubectl -
安装好
kubernetes集群 -
下载
velero客户端-
Mac可直接使用HomeBrew进行安装
brew install velero -
Linux从Github的Releases页面下载，进行解压安装
tar -xvf <RELEASE-TARBALL-NAME>.tar.gz -
Windows使用
Chocolatey进行安装choco install velero -
通用安装方式，在Release页面下载对应平台的压缩包，解压到PATH环境变量下。
-
三、部署
这里以腾讯云cos为例
1.创建存储桶，需要开通对象存储cos服务
2.创建凭据
这里的AK和SK登录腾讯云，通过【访问管理】-【访问密钥】-【API密钥管理】进行查看
1vim credentials-velero
2[default]
3aws_access_key_id = xxx
4aws_secret_access_key = xxx
3.安装脚本
1vim velero-install.sh
2velero install \
3 --provider aws \
4 --bucket bucket-name \
5 --prefix "backup" \
6 --namespace velero \
7 --secret-file ./credentials-velero \
8 --velero-pod-cpu-request 200m \
9 --velero-pod-mem-request 200Mi \
10 --velero-pod-cpu-limit 200m \
11 --velero-pod-mem-limit 200Mi \
12 --use-volume-snapshots=false \
13 --use-restic \
14 --restic-pod-cpu-request 200m \
15 --restic-pod-mem-request 200Mi \
16 --restic-pod-cpu-limit 200m \
17 --restic-pod-mem-limit 200Mi \
18 --plugins velero/velero-plugin-for-aws:v1.1.0 \
19 --backup-location-config region=区域,s3ForcePathStyle="false",s3Url=http://cos.区域.myqcloud.com
4.执行安装
1./velero-install.sh
2CustomResourceDefinition/backups.velero.io: attempting to create resource
3CustomResourceDefinition/backups.velero.io: created
4CustomResourceDefinition/backupstoragelocations.velero.io: attempting to create resource
5CustomResourceDefinition/backupstoragelocations.velero.io: created
6CustomResourceDefinition/deletebackuprequests.velero.io: attempting to create resource
7CustomResourceDefinition/deletebackuprequests.velero.io: created
8CustomResourceDefinition/downloadrequests.velero.io: attempting to create resource
9CustomResourceDefinition/downloadrequests.velero.io: created
10CustomResourceDefinition/podvolumebackups.velero.io: attempting to create resource
11CustomResourceDefinition/podvolumebackups.velero.io: created
12CustomResourceDefinition/podvolumerestores.velero.io: attempting to create resource
13CustomResourceDefinition/podvolumerestores.velero.io: created
14CustomResourceDefinition/resticrepositories.velero.io: attempting to create resource
15CustomResourceDefinition/resticrepositories.velero.io: created
16CustomResourceDefinition/restores.velero.io: attempting to create resource
17CustomResourceDefinition/restores.velero.io: created
18CustomResourceDefinition/schedules.velero.io: attempting to create resource
19CustomResourceDefinition/schedules.velero.io: created
20CustomResourceDefinition/serverstatusrequests.velero.io: attempting to create resource
21CustomResourceDefinition/serverstatusrequests.velero.io: created
22CustomResourceDefinition/volumesnapshotlocations.velero.io: attempting to create resource
23CustomResourceDefinition/volumesnapshotlocations.velero.io: created
24Waiting for resources to be ready in cluster...
25Namespace/velero: attempting to create resource
26Namespace/velero: created
27ClusterRoleBinding/velero: attempting to create resource
28ClusterRoleBinding/velero: created
29ServiceAccount/velero: attempting to create resource
30ServiceAccount/velero: created
31Secret/cloud-credentials: attempting to create resource
32Secret/cloud-credentials: created
33BackupStorageLocation/default: attempting to create resource
34BackupStorageLocation/default: created
35Deployment/velero: attempting to create resource
36Deployment/velero: created
37DaemonSet/restic: attempting to create resource
38DaemonSet/restic: created
39Velero is installed! ⛵ Use 'kubectl logs deployment/velero -n velero' to view the status.
5 查看
1kubectl get po -n velero
2NAME READY STATUS RESTARTS AGE
3restic-f5jk8 1/1 Running 0 8s
4restic-kk7xj 1/1 Running 0 29s
5restic-qddwg 1/1 Running 0 38s
6restic-vn9gk 1/1 Running 0 35s
7velero-8fb9578d5-xkqbv 1/1 Running 0 39s
四、使用
1.创建备份
1velero create backup cloud-industry --include-namespaces cloud-industry
2Backup request "cloud-industry" submitted successfully.
3Run `velero backup describe cloud-industry` or `velero backup logs cloud-industry` for more details.
2.查看详情
1velero backup describe cloud-industry
2Name: cloud-industry
3Namespace: velero
4Labels: velero.io/storage-location=default
5Annotations: velero.io/source-cluster-k8s-gitversion=v1.18.2
6 velero.io/source-cluster-k8s-major-version=1
7 velero.io/source-cluster-k8s-minor-version=18
8
9Phase: Completed
10
11Errors: 0
12Warnings: 0
13
14Namespaces:
15 Included: cloud-industry
16 Excluded: <none>
17
18Resources:
19 Included: *
20 Excluded: <none>
21 Cluster-scoped: auto
22
23Label selector: <none>
24
25Storage Location: default
26
27Velero-Native Snapshot PVs: auto
28
29TTL: 720h0m0s
30
31Hooks: <none>
32
33Backup Format Version: 1
34
35Started: 2020-09-09 22:32:47 +0800 CST
36Completed: 2020-09-09 22:33:19 +0800 CST
37
38Expiration: 2020-10-09 22:32:47 +0800 CST
39
40Total items to be backed up: 2305
41Items backed up: 2305
42
43Velero-Native Snapshots: <none included>
3.查看备份
- 使用velero命令查看
1velero backup get
2NAME STATUS ERRORS WARNINGS CREATED EXPIRES STORAGE LOCATION SELECTOR
3cloud-industry Completed 0 0 2020-09-09 22:32:47 +0800 CST 29d default <none>
- 使用s3cmd查看
1s3cmd ls s3://yaml-backup-1257948216/backup/backups/cloud-industry/
22020-09-09 14:33 29 s3://bucket-name-1257948216/poc-h/backups/cloud-industry/cloud-industry-csi-volumesnapshotcontents.json.gz
32020-09-09 14:33 29 s3://bucket-name-1257948216/backup/backups/cloud-industry/cloud-industry-csi-volumesnapshots.json.gz
42020-09-09 14:33 55047 s3://bucket-name-1257948216/backup/backups/cloud-industry/cloud-industry-logs.gz
52020-09-09 14:33 29 s3://bucket-name-1257948216/backup/backups/cloud-industry/cloud-industry-podvolumebackups.json.gz
62020-09-09 14:33 20424 s3://bucket-name-1257948216/backup/backups/cloud-industry/cloud-industry-resource-list.json.gz
72020-09-09 14:33 29 s3://bucket-name-1257948216/backup/backups/cloud-industry/cloud-industry-volumesnapshots.json.gz
82020-09-09 14:33 368225 s3://bucket-name-1257948216/backup/backups/cloud-industry/cloud-industry.tar.gz
92020-09-09 14:33 1108 s3://bucket-name-1257948216/backup/backups/cloud-industry/velero-backup.json
4.还原
1velero restore create cloud-industry --from-backup cloud-industry
2Restore request "cloud-industry" submitted successfully.
3Run `velero restore describe cloud-industry` or `velero restore logs cloud-industry` for more details.
五、高级用法
上面已经实现了手动备份和还原功能，下面我们schedule实现自动备份功能。
1.排除要备份的名称空间
1kubectl label -n kube-system namespace/kube-system velero.io/exclude-from-backup=true
2namespace/kube-system labeled
3kubectl label -n kube-public namespace/kube-public velero.io/exclude-from-backup=true
4namespace/kube-public labeled
5kubectl label -n default namespace/default velero.io/exclude-from-backup=true
6namespace/default labeled
2.使用schedule设置定时备份
1# 每天2点进行备份，保留最近3天的备份
2velero create schedule auto-backup --schedule="0 2 * * *" --ttl 72h
3Schedule "auto-backup" created successfully.
3.其他定时备份的示例
1# 每日1点进行备份
2velero create schedule <SCHEDULE NAME> --schedule="0 1 * * *"
3# 每日2点进行备份，备份保留72小时
4velero create schedule <SCHEDULE NAME> --schedule="0 2 * * *" --ttl 72h
5# 每6小时进行一次备份
6velero create schedule <SCHEDULE NAME> --schedule="@every 6h"
7# 每日对 web namespace 进行一次备份
8velero create schedule <SCHEDULE NAME> --schedule="@every 24h" --include-namespaces web
备注
定时备份的名称为：<SCHEDULE NAME>-<TIMESTAMP>，恢复命令为：velero restore create --from-backup <SCHEDULE NAME>-<TIMESTAMP>。
4.查看定时备份
1velero schedule get
2NAME STATUS CREATED SCHEDULE BACKUP TTL LAST BACKUP SELECTOR
3auto-backup Enabled 2020-09-09 23:12:09 +0800 CST 0 2 * * * 72h0m0s 1m ago <none>
六、参考
- 原文作者：黄忠德
- 原文链接：https://huangzhongde.cn/post/Kubernetes/Using_velero_manage_kubernetes_resources/
- 版权声明：本作品采用知识共享署名-非商业性使用-禁止演绎 4.0 国际许可协议进行许可，非商业转载请注明出处（作者，原文链接），商业转载请联系作者获得授权。