gitlab服务有非常简洁的每日备份命令，
从production的gitlab的每日备份中restore到backup环境也非常方便。
一、Production gitlab每日备份
1. Production gitlab环境上编写脚本
cat /root/gitlab_bak.sh
gitlab-rake gitlab:backup:create > /var/opt/gitlab/backups/log/$(date +"%Y-%m-%d-%H:%M:%S").log
运行这个脚本，会生成tar包到gitlab配置文件中指定的backup路径。
gitlab配置文件（/etc/gitlab/gitlab.rb）中配置backup路径的部分如下：（默认路径为/var/opt/gitlab/backups）
### Backup Settings
###! Docs: https://docs.gitlab.com/omnibus/settings/backups.html
# gitlab_rails['manage_backup_path'] = true
gitlab_rails['backup_path'] = "/var/opt/gitlab/backups"
2. 创建定时任务进行每日备份
0 2 * * * /root/gitlab_bak.sh
二、Backup的gitlab从每日备份中restore
1. Backup gitlab环境上创建脚本
restore脚本如下：
cat gitlab_restore.sh
#!/bin/bash
backup_file=$(ls -t /var/opt/gitlab/backups/*.tar|head -1)
restore_path=/var/opt/gitlab/backups
echo "$(date "+%Y-%m-%d %H:%M:%S") Start restoring from backup file: $backup_file"
# stop db related services
gitlab-ctl stop unicorn
gitlab-ctl stop sidekiq
# prepare the backupfiles to restore path
rm -rf ${restore_path}/*
cp ${backup_file} ${restore_path}/
chmod 777 -R ${restore_path}/*
gitlab-rake gitlab:backup:restore force=yes
gitlab-ctl reconfigure && gitlab-ctl restart
rm -rf ${restore_path}/*
echo "$(date "+%Y-%m-%d %H:%M:%S") End restoring, backup gitlab url: http://xxx"
需要把production备份的/var/opt/gitlab/backups/*.tar 拷贝到backup环境。
2. 运行脚本即可完成restore
注：
注意定期清理路径：/var/opt/gitlab/git-data/repositories/+gitaly/tmp
这个路径会占比较大的空间。可能会导致空间不足，restore失败。