基于KubeSphere在K8S发布微服务ruoyi----6.部署基础环境harbor私有仓库
1 下载并解压
cd /tmp
wget https://github.com/goharbor/harbor/releases/download/v2.6.0/harbor-offline-installer-v2.6.0.tgz
tar xf harbor-offline-installer-v2.6.0.tgz -C /opt/
cd /opt/
mv harbor/ harbor-v2.6.0
ln -s /opt/harbor-v2.6.0/ /opt/harbor
2 编辑配置文件
cp harbor.yml.tmpl harbor.yml
# vi /opt/harbor/harbor.yml
# 以下是修改项,手动在配置文件中更改
hostname: harbor.st.com
http:
port: 81
harbor_admin_password:Harbor12345
data_volume: /data/harbor
log:
level: info
rotate_count: 50
rotate_size:200M
location: /data/harbor/logs
#默认port:80,改成81,为了避免和gitlab的80端口产生冲突
https需注释掉,否则无法启动
创建日志目录
mkdir -p /data/harbor/logs
检查配置:
cat /opt/harbor/harbor.yml| egrep -v "#|^$"
3 使用docker-compose启动harbor harbor依赖compose做单机编排
cd /opt/harbor/
yum install docker-compose -y 需要epel源
epel源
wget -O /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repo
sh /opt/harbor/install.sh
docker-compose ps
docker ps -a
4 使用dns服务器或host文件解析harbor
追加解析
echo 192.168.1.143 harbor.st.com>>/etc/hosts
5.insecure-registries选项修改,允许使用http登录
Harbor客户端设置示例登录前客户端先设置好insecure-registries
Worker节点之前改过了,使用了阿里云加速配置
cat /etc/docker/daemon.json
{
"registry-mirrors": ["https://82m9ar63.mirror.aliyuncs.com"],
"exec-opts": ["native.cgroupdriver=systemd"],
"log-driver": "json-file",
"insecure-registries": ["registry.access.redhat.com","quay.io","harbor.st.com:81","192.168.1.143:81"],
"log-opts": {
"max-size": "100m"
},
"storage-driver": "overlay2"
}
Master节点未改过,未使用阿里云加速配置
{
"log-opts": {
"max-size": "5m",
"max-file":"3"
},
"exec-opts": ["native.cgroupdriver=systemd"],
"insecure-registries": ["registry.access.redhat.com","quay.io","harbor.st.com:81","192.168.1.143:81"]
}
重启docker
systemctl restart docker
6.登录测试
修改了默认端口80(已被Gitlab占用),改用了81端口
浏览器输入:harbor.st.com:81或http://192.168.1.143:81/
用户名:admin 密码:Harbor12345
docker登录
docker login harbor.st.com -uadmin -pHarbor12345
排错: Harbor访问出现502Bad Gateway的情况
cd /opt/harbor/
使用docker-compose ps [列出所有容器] 查看容器情况
入调试模式查看异常情况
docker-compose up [列出所有容器的日志,退出时关闭容器]
docker-compose up & 后台执行
docker ps 检查是否有unhealthy的容器
解决方法:
重启unhealthy的容器
重启docker服务
重启计算机并执行docker-compose up
放大招:重新安装 /opt/harbor/install.sh
data_volume已存在本机,重新编排后以前上传的镜像仍存在