极狐GitLab 如何在 helm 中恢复数据

极小狐2024-04-14 20:07

本文作者：徐晓伟

GitLab是一个全球知名的一体化 DevOps 平台，很多人都通过私有化部署 GitLab 来进行源代码托管。极狐GitLab 是GitLab 在中国的发行版，专门为中国程序员服务。可以一键式部署极狐GitLab。

本文主要讲述了如何在极狐GitLab 恢复 GitLab 历史备份数据

文档

  1. 恢复 极狐GitLab 安装实例

本实例数据恢复

如果当前 极狐GitLab 实例数据被误删、实例异常，可通过历史备份进行数据恢复

记下数据库客户端的当前副本数，供后续重启使用

复制代码 
# -n gitlab-test：代表命名空间
# release=my-gitlab：代表 helm 安装 gitlab 时的名称
# -lapp=sidekiq：代表 sidekiq-all-in-1-v2 Deployment，默认 1 个副本
# -lapp=webservice：代表 webservice-default Deployment，默认 2 个副本
# -lapp=prometheus：代表 prometheus-server Deployment，默认 1 个副本
kubectl -n gitlab-test get deploy -lapp=sidekiq,release=my-gitlab -o jsonpath='{.items[].spec.replicas}{"\n"}'
kubectl -n gitlab-test get deploy -lapp=webservice,release=my-gitlab -o jsonpath='{.items[].spec.replicas}{"\n"}'
kubectl -n gitlab-test get deploy -lapp=prometheus,release=my-gitlab -o jsonpath='{.items[].spec.replicas}{"\n"}'

停止数据库的客户端，防止锁干扰恢复过程

复制代码 
kubectl -n gitlab-test scale deploy -lapp=sidekiq,release=my-gitlab --replicas=0
kubectl -n gitlab-test scale deploy -lapp=webservice,release=my-gitlab --replicas=0
kubectl -n gitlab-test scale deploy -lapp=prometheus,release=my-gitlab --replicas=0

查看 Toolbox pod

复制代码 
kubectl -n gitlab-test get pods -lrelease=my-gitlab,app=toolbox

# manual-backup-1-xxx：代表个人备份的 pod，无需关注
# my-gitlab-toolbox-backup-xxx-xxx：代表定时任务备份的 pod，无需关注
# my-gitlab-toolbox-xxx-xxx：代表 Toolbox pod 容器，需要记录此名称，用于后续使用
[root@anolis-7-9 ~]# kubectl -n gitlab-test get pods -lrelease=my-gitlab,app=toolbox
NAME                                      READY   STATUS      RESTARTS   AGE
manual-backup-1-9db78                     0/1     Completed   0          46h
my-gitlab-toolbox-5b7c4f8c6b-zccrv        1/1     Running     0          8m46s
my-gitlab-toolbox-backup-28393980-f9m8h   0/1     Completed   0          3h3m
[root@anolis-7-9 ~]#

直接使用备份数据的 URL 链接恢复数据（推荐）

备份数据目录
复制代码 
[root@anolis-7-9 ~]# ll -h /gitlab-test/my-gitlab-minio-pv/gitlab-backups/
total 11M
-rw-rw-r-- 1 1000 1000 5.1M Dec 25 14:18 1703484999_2023_12_25_16.7.0-ee_gitlab_backup.tar
-rw-r--r-- 1 1000 1000 5.1M Dec 27 09:38 1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar
[root@anolis-7-9 ~]#
使用 httpd 创建静态资源服务

使用其他方式创建静态资源服务也可以

复制代码 
docker run \
  --restart=always \
  -itd \
  --privileged=true \
  --name 2.4.48 \
  -p 180:80 \
  -v /gitlab-test/my-gitlab-minio-pv/gitlab-backups/:/usr/local/apache2/htdocs/ \
  -v /etc/localtime:/etc/localtime \
  -d httpd:2.4.48
查看静态资源

  • 访问 httpd

使用静态资源 URL 恢复数据
复制代码 
kubectl -n gitlab-test exec my-gitlab-toolbox-5b7c4f8c6b-zccrv -it -- backup-utility --restore -f http://172.25.25.32:180/1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar

成功示例

复制代码 
[root@anolis-7-9 ~]# kubectl -n gitlab-test exec my-gitlab-toolbox-5b7c4f8c6b-zccrv -it -- backup-utility --restore -f http://172.25.25.32:180/1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar
Defaulted container "toolbox" out of: toolbox, certificates (init), configure (init)
Downloading from http://172.25.25.32:180/1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar
############################################################################################################################################################ 100.0%
Unpacking backup
2023-12-27 12:28:10 +0800 -- Restoring database ... 
2023-12-27 12:28:10 +0800 -- Be sure to stop Puma, Sidekiq, and any other process that
connects to the database before proceeding. For Omnibus
installs, see the following link for more information:
https://docs.gitlab.com/ee/raketasks/backup_restore.html#restore-for-omnibus-gitlab-installations

Before restoring the database, we will remove all existing
tables to avoid future upgrade problems. Be aware that if you have
custom tables in the GitLab database these tables and all data will be
removed.
Do you want to continue (yes/no)? yes
Removing all tables. Press `Ctrl-C` within 5 seconds to abort
2023-12-27 12:28:32 +0800 -- Cleaning the database ... 
2023-12-27 12:28:35 +0800 -- done
Restoring PostgreSQL database gitlabhq_production ... ERROR:  must be owner of extension pg_trgm
ERROR:  must be owner of extension btree_gist
ERROR:  must be owner of extension pg_trgm
SET
 set_config 
------------
...
[DONE]
Source backup for the database ci doesn't exist. Skipping the task
2023-12-27 12:28:53 +0800 -- Restoring database ... done
2023-12-27 12:28:53 +0800 -- Deleting backup and restore PID file ... done
2023-12-27 12:29:08 +0800 -- Restoring repositories ... 
{"command":"restore","gl_project_path":"xuxiaowei-com-cn.wiki","level":"info","msg":"started restore","pid":65,"relative_path":"@groups/19/58/19581e27de7ced00ff1ce50b2047e7a567c76b1cbaebabe5ef03f7c3017bb5b7.wiki.git","storage_name":"default","time":"2023-12-27T04:29:08.837Z"}
...
2023-12-27 12:29:09 +0800 -- Restoring repositories ... done
2023-12-27 12:29:09 +0800 -- Deleting backup and restore PID file ... done
Restoring artifacts ...
done
Restoring packages ...
Restoring pages ...
[root@anolis-7-9 ~]#
恢复副本数量

  • 使用上述命令记录的副本数进行副本恢复

    kubectl -n gitlab-test scale deploy -lapp=sidekiq,release=my-gitlab --replicas=1
    kubectl -n gitlab-test scale deploy -lapp=webservice,release=my-gitlab --replicas=2
    kubectl -n gitlab-test scale deploy -lapp=prometheus,release=my-gitlab --replicas=1

复制备份数据到容器 Toolbox pod 内进行数据恢复（不推荐）

不推荐的原因：如果备份数据过大（例如：超过 10G），可能会存在没有复制完成的情况，导致无法解压备份数据，无法恢复

复制备份数据到容器 Toolbox pod 内

  1. 查看历史备份数据

    复制代码 
    [root@anolis-7-9 ~]# ll -h /gitlab-test/my-gitlab-minio-pv/gitlab-backups/
total 11M
-rw-rw-r-- 1 1000 1000 5.1M Dec 25 14:18 1703484999_2023_12_25_16.7.0-ee_gitlab_backup.tar
-rw-r--r-- 1 1000 1000 5.1M Dec 27 09:38 1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar
[root@anolis-7-9 ~]#

  2. 复制到容器 Toolbox pod 内的 /srv/gitlab/tmp/backups 目录

    复制代码 
    [root@anolis-7-9 ~]# kubectl -n gitlab-test cp /gitlab-test/my-gitlab-minio-pv/gitlab-backups/1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar my-gitlab-toolbox-5b7c4f8c6b-zccrv:/srv/gitlab/tmp/backups
Defaulted container "toolbox" out of: toolbox, certificates (init), configure (init)
[root@anolis-7-9 ~]# 
[root@anolis-7-9 ~]# kubectl -n gitlab-test exec -it my-gitlab-toolbox-5b7c4f8c6b-zccrv ls /srv/gitlab/tmp/backups
kubectl exec [POD] [COMMAND] is DEPRECATED and will be removed in a future version. Use kubectl exec [POD] -- [COMMAND] instead.
Defaulted container "toolbox" out of: toolbox, certificates (init), configure (init)
1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar  gitlab-shell
[root@anolis-7-9 ~]#
恢复数据

  • 使用 容器内 /srv/gitlab/tmp/backups 目录下的部分文件名， 格式：<timestamp>_<version>

    复制代码 
    kubectl -n gitlab-test exec my-gitlab-toolbox-5b7c4f8c6b-zccrv -it -- backup-utility --restore -t 1703640824_2023_12_27_16.7.0-ee

  • 也可以使用 file:///<path> 指定文件路径绝对路径

    复制代码 
    kubectl -n gitlab-test exec my-gitlab-toolbox-5b7c4f8c6b-zccrv -it -- backup-utility --restore -t file:///home/git/1703640824_2023_12_27_16.7.0-ee_gitlab_backup.tar

成功示例

复制代码 
[root@anolis-7-9 ~]# kubectl -n gitlab-test exec my-gitlab-toolbox-5b7c4f8c6b-zccrv -it -- backup-utility --restore -t 1703640824_2023_12_27_16.7.0-ee
Unpacking backup
2023-12-27 12:38:22 +0800 -- Restoring database ... 
2023-12-27 12:38:22 +0800 -- Be sure to stop Puma, Sidekiq, and any other process that
connects to the database before proceeding. For Omnibus
installs, see the following link for more information:
https://docs.gitlab.com/ee/raketasks/backup_restore.html#restore-for-omnibus-gitlab-installations

Before restoring the database, we will remove all existing
tables to avoid future upgrade problems. Be aware that if you have
custom tables in the GitLab database these tables and all data will be
removed.
Do you want to continue (yes/no)? yes
Removing all tables. Press `Ctrl-C` within 5 seconds to abort
2023-12-27 12:41:21 +0800 -- Cleaning the database ... 
2023-12-27 12:41:23 +0800 -- done
Restoring PostgreSQL database gitlabhq_production ... ERROR:  must be owner of extension pg_trgm
ERROR:  must be owner of extension btree_gist
ERROR:  must be owner of extension pg_trgm
SET
 set_config 
------------
...
[DONE]
Source backup for the database ci doesn't exist. Skipping the task
2023-12-27 12:41:37 +0800 -- Restoring database ... done
2023-12-27 12:41:37 +0800 -- Deleting backup and restore PID file ... done
2023-12-27 12:41:52 +0800 -- Restoring repositories ... 
{"command":"restore","gl_project_path":"xuxiaowei-com-cn.wiki","level":"info","msg":"started restore","pid":236,"relative_path":"@groups/19/58/19581e27de7ced00ff1ce50b2047e7a567c76b1cbaebabe5ef03f7c3017bb5b7.wiki.git","storage_name":"default","time":"2023-12-27T04:41:52.584Z"}
...
2023-12-27 12:41:53 +0800 -- Restoring repositories ... done
2023-12-27 12:41:53 +0800 -- Deleting backup and restore PID file ... done
Restoring artifacts ...
done
Restoring packages ...
Restoring pages ...
[root@anolis-7-9 ~]#
恢复副本数量

  • 使用上述命令记录的副本数进行副本恢复

    kubectl -n gitlab-test scale deploy -lapp=sidekiq,release=my-gitlab --replicas=1
    kubectl -n gitlab-test scale deploy -lapp=webservice,release=my-gitlab --replicas=2
    kubectl -n gitlab-test scale deploy -lapp=prometheus,release=my-gitlab --replicas=1

问题

自定义 toolbox Deployment hostAliases

问题
复制代码 
2023-12-27 11:48:42 +0800 -- Restoring repositories ... done
2023-12-27 11:48:42 +0800 -- Deleting backup and restore PID file ... done
Restoring artifacts ...
[Error] ERROR: [Errno -2] Name or service not known
ERROR: Connection Error: Error resolving a server hostname.
Please check the servers address specified in 'host_base', 'host_bucket', 'cloudfront_host', 'website_endpoint'

sync existing of artifacts failed
command terminated with exit code 1
解决方案
复制代码 
kubectl -n gitlab-test edit deployments.apps my-gitlab-toolbox

      # 添加 MinIO 对象储存的 host 解析
      hostAliases:
        - hostnames:
            - minio.test.helm.xuxiaowei.cn
          ip: 172.25.25.32

信任 MinIO 对象储存域名证书

问题
复制代码 
Restoring artifacts ...
[Error] ERROR: SSL certificate verification failure: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get local issuer certificate (_ssl.c:1129)

sync existing of artifacts failed
command terminated with exit code 1
解决办法
复制代码 
kubectl -n gitlab-test edit deployments.apps my-gitlab-toolbox

      containers:
        # 不验证证书
        - args:
            - /bin/bash
            - -c
            - cp -v -r -L /etc/gitlab/.s3cfg $HOME/.s3cfg && echo "check_ssl_certificate=false" >> $HOME/.s3cfg && while sleep 3600; do :; done

号外号外！

极狐GitLab 正在推出DevSecOps 成熟度测评！链接：https://gitlab.cn/devsecops-assessment/ 测评非常全面并提供了可靠建议，即使不付费买产品，对自己想要落地 DevSecOps 的用户具有很高的参考意义！快来动手试试吧！

相关推荐
老华带你飞1 小时前
口腔助手|口腔挂号预约小程序|基于微信小程序的口腔门诊预约系统的设计与实现(源码+数据库+文档)
java·数据库·微信小程序·小程序·论文·毕设·口腔小程序
hqxstudying1 小时前
J2EE模式---服务层模式
java·数据库·后端·spring·oracle·java-ee
Yu_Lijing1 小时前
MySQL进阶学习与初阶复习第四天
数据库·学习·mysql
大熊程序猿1 小时前
net8.0一键创建支持(Orm-Sqlite-MySql-SqlServer)
数据库·mysql·sqlite
-SGlow-10 小时前
MySQL相关概念和易错知识点（2）（表结构的操作、数据类型、约束）
linux·运维·服务器·数据库·mysql
明月56611 小时前
Oracle 误删数据恢复
数据库·oracle
♡喜欢做梦13 小时前
【MySQL】深入浅出事务：保证数据一致性的核心武器
数据库·mysql
遇见你的雩风13 小时前
MySQL的认识与基本操作
数据库·mysql
dblens 数据库管理和开发工具13 小时前
MySQL新增字段DDL：锁表全解析、避坑指南与实战案例
数据库·mysql·dblens·dblens mysql·数据库连接管理
weixin_4196583113 小时前
MySQL的基础操作
数据库·mysql
热门推荐
01Qwen3-Coder 快速上手教程 | Qwen Code + Claude Code02vue数据变化但页面不变03全球最强模型Grok4，国内已可免费使用！（附教程）04扣子开源本地部署教程 丨Coze智能体小白喂饭级指南05KGG转MP3工具|非KGM文件|解密音频06sqli-labs 靶场 less-8、9、10 第八关到第十关详解：布尔注入，时间注入07干翻 Typora！MilkUp：完全免费的桌面端 Markdown 编辑器！08Coze 开源了，送上保姆级私有化部署方案【建议收藏】09【2025.7.18】更新vscode后所有.vue文件template标签后报红的临时解决办法，Vue - Official 插件3.0.2导致1001-开源版COZE-字节 Coze Studio 重磅开源！保姆级本地安装教程，手把手带你体验