配置Sinopia服务器

linux系统配置

• 网络环境配置
• 开放端口
  默认只开放22端口

Node环境安装

• 将下载的linux版本的node放到/opt目录中
• 如果安装包以xz结尾使用xz -d *.tar.xz命令解压
• 将node、npm映射到环境变量中

配置Sinopia

• 详细配置信息
• 后台运行服务命令 nohup sinopia &

上传缓存包

在有网络环境的机器上用npm连接有上级服务器的私服下载需要的包
注意：下载前确保本地无缓存，可以使用npm cache clean命令清除缓存包然后再安装
将下载的包上传到sinopia目录下

---

内网npm服务器特殊配置

storage: ./storage

auth:
  htpasswd:
    file: ./htpasswd
    # Maximum amount of users allowed to register, defaults to "+inf".
    # You can set this to -1 to disable registration.
    max_users: -1

# a list of other known repositories we can talk to
# 去掉上级链接
# uplinks:
  # npmjs:
    # url: https://registry.npmjs.org/
    # url: http://registry.npm.taobao.org/

packages:
  '@*/*':
    # scoped packages
    access: $all
    publish: $authenticated

  '*':
    # allow all users (including non-authenticated users) to read and
    # publish all packages
    #
    # you can specify usernames/groupnames (depending on your auth plugin)
    # and three keywords: "$all", "$anonymous", "$authenticated"
    access: $all

    # allow all known users to publish packages
    # (anyone can register by default, remember?)
    publish: $authenticated

    # if package is not available locally, proxy requests to 'npmjs' registry
    # 去掉代理下载
    # proxy: npmjs

# log settings
logs:
  - {type: stdout, format: pretty, level: http}
  #- {type: file, path: sinopia.log, level: info}

listen: 0.0.0.0:4873

补充编辑（2017-10-31）：
sinopia突然有段时间经常在下载包的时候出现offline情况，解决方案，延长错误超时时间。

uplinks:
  npmjs:
    url: https://registry.npmjs.org/
    #url: http://registry.npm.taobao.org/
    #url: http://r.cnpmjs.org/
    #url: https://registry.nodejitsu.com/
    #url: https://skimdb.npmjs.com/registry/

    #设置请求无应答超时时间
    # amount of time to wait for repository to respond
    # before giving up and use the local cached copy
    timeout: 20s

    #设置数据认为最新的时间为2分钟，2分钟同一个数据的请求不会向上游服务器请求
    # maximum time in which data is considered up to date
    #
    # default is 2 minutes, so server won't request the same data from
    # uplink if a similar request was made less than 2 minutes ago
    maxage: 2m

    #设置访问失败达到某次数，就停止一段时间不访问上游服务器，默认是2次不应答，5分钟不去请求
    # if two subsequent requests fail, no further requests will be sent to
    # this uplink for five minutes
    max_fails: 20000
    fail_timeout: 50m

    # timeouts are defined in the same way as nginx, see:
    # http://wiki.nginx.org/ConfigNotation

---

附加链接：
http://blog.csdn.net/u010130282/article/details/51892625
https://my.oschina.net/blogshi/blog/260953
http://www.cnblogs.com/xinjie10001/p/6295020.html
http://www.cnblogs.com/laov/p/3541414.html