工具链备份
背景
-
其实在之前自己备份意识是不够的，而且我发现公司所有工具链都在阿里云外网访问，更可怕的是没有定时备份(docker 启动并没有持久化)。不敢想象被***会是什么结果。我到公司后第一件事就是迁入内网，但是内网机器好多人都有账号而且sudo权限也大，所以备份也是要做的，毕竟内网也不安全。
- 备份作为最基础的运维是必要的，备份也是提高服务可用性中重要的一环，在灾难来临之时，可以做到快速恢复服务。举个例子警示一下，比如gitlab误删数据事故 当然gitlab有多重备份机制，恢复是肯定的。想一想如果gitlab没有备份这将会是巨大灾难，过往种种事件都在警示我们做好备份，只要数据还在一切都不是问题。
环境
- confluence,jira,gitlab,yapi都是采用docker启动
- gitlab，yapi做了持久化
- 尴尬的一点是jira和confluence没有做持久化，所以备份是必须的，有空这个坑得处理一下，因为下面做的备份是按天，如果容器崩溃会丢失一部分数据。
重要提示：容器做数据持久化
gitlab备份
cat /root/backup_gitlab.sh 备份脚本
#!/bin/bash
backup_time=`date +"%Y-%m-%d"`
docker exec -t gitlab-docker gitlab-rake gitlab:backup:create
if [ $? -eq 0 ]
then
echo "gitlab ${backup_time} Successful backup" >> /root/backup.log
else
echo "gitlab ${backup_time} backup failed" >> /root/backup.log
fi
find /mnt/md0/gitlab-docker/data/backups/ -maxdepth 1 -mtime +7 -name "*.tar" | xargs rm -f
设置定时任务
30 1 * * * /root/backup_gitlab.sh >/dev/null 2>&1jira备份
root@pc-Super-Server:~# cat backup_jira.sh
#!/bin/bash
backup_time=`date +"%Y-%m-%d"`
PGPASSWORD=jiradb pg_dump -h 172.168.1.235 -p 9999 -U jiradb > /mnt/md0/backup_data/jira/jiradb_${backup_time}.sql
if [ $? -eq 0 ]
then
echo "jira ${backup_time} sql Successful backup" >> /root/backup.log
else
echo "jira ${backup_time} sql backup failed" >> /root/backup.log
fi
docker cp jira:/var/jira/data/ /mnt/md0/backup_data/jira/data/
docker cp jira:/var/jira/dbconfig.xml /mnt/md0/backup_data/jira/
docker cp jira:/var/jira/caches/ /mnt/md0/backup_data/jira/
# 查找7天前的文件执行删除
find /mnt/md0/backup_data/jira/ -maxdepth 1 -mtime +7 -name "*.sql" | xargs rm -f
30 3 * * * /root/backup_jira.sh >/dev/null 2>&1
confluence备份
root@pc-Super-Server:~# cat backup_confluence.sh
#!/bin/bash
backup_time=`date +"%Y-%m-%d"`
PGPASSWORD=confdb pg_dump -h 172.168.1.235 -p 9999 -U confdb > /mnt/md0/backup_data/confluence/confdb_${backup_time}.sql
if [ $? -eq 0 ]
then
echo "confluence ${backup_time} sql Successful backup" >> /root/backup.log
else
echo "confluence ${backup_time} sql backup failed" >> /root/backup.log
fi
docker cp confluence:/var/atlassian/confluence/attachments /mnt/md0/backup_data/confluence/
docker cp confluence:/var/atlassian/confluence/confluence.cfg.xml /mnt/md0/backup_data/confluence/
docker cp confluence:/var/atlassian/confluence/index /mnt/md0/backup_data/confluence/
# 查找7天前的文件执行删除
find /mnt/md0/backup_data/confluence/ -maxdepth 1 -mtime +7 -name "*.sql" | xargs rm -f
0 3 * * * /root/backup_confluence.sh >/dev/null 2>&1
yapi备份
root@pc-Super-Server:~# cat backup_yapi.sh 备份脚本
#!/bin/bash
backup_time=`date +"%Y-%m-%d"`
docker exec -t mongo-yapi mongodump -u yapi -p yapi2019 -d yapi -o /data/db/backups/`date +'%Y-%m-%d'`/
if [ $? -eq 0 ]
then
echo "yapi ${backup_time} Successful backup" >> /root/backup.log
else
echo "yapi ${backup_time} backup failed" >> /root/backup.log
fi
find /data/yapi-data/backups/ -maxdepth 1 -type d -mtime +7 |xargs rm -rf
crontab 定时任务
30 2 * * * /root/backup_yapi.sh >/dev/null 2>&1
yapi恢复
- docker 启动一个全新的yapi 和mongo
mongo
use admin
db.createUser(
{ user: "admin",
pwd: "admin",
roles: [ { role: "userAdminAnyDatabase", db: "admin" } ]
}
)
db.createUser(
{ user: "yapi",
pwd: "yapi2019",
roles: [ { role: "dbAdmin", db: "yapi" } ]
}
)
exit
mongorestore -u yapi -p yapi2019 -d yapi --authenticationDatabase admin --drop --dir /root/yapi/
/root/yapi 是备份数据目录