README.html



<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
  <meta charset="utf-8">
  
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  
  <title>FileManager设计文档 &mdash; PaddlePaddle  documentation</title>
  

    <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
  

        <link rel="index" title="Index"
              href="../../genindex.html"/>
        <link rel="search" title="Search" href="../../search.html"/>
    <link rel="top" title="PaddlePaddle  documentation" href="../../index.html"/> 

  
  <script src="../../_static/js/modernizr.min.js"></script>

</head>

<body class="wy-body-for-nav" role="document">

  <div class="wy-grid-for-nav">

    
    <nav data-toggle="wy-nav-shift" class="wy-nav-side">
      <div class="wy-side-scroll">
        <div class="wy-side-nav-search">
          

            <a href="../../index_en.html" class="icon icon-home"> PaddlePaddle
          

          </a>

          
<div role="search">
  <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" />
    <input type="hidden" name="check_keywords" value="yes" />
    <input type="hidden" name="area" value="default" />
  </form>
</div>

          
        </div>

        <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
          
            
                <ul>
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_en.html">GET STARTED</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../build_and_install/index_en.html">Install and Build</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../howto/index_en.html">HOW TO</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../dev/index_en.html">Development</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_en.html">FAQ</a></li>
</ul>

            
        </div>
      </div>
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">

      
      <nav class="wy-nav-top" role="navigation" aria-label="top navigation">
        <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
        <a href="../../index_en.html">PaddlePaddle</a>
      </nav>


      <div class="wy-nav-content">
        <div class="rst-content">
          

<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
    <li><a href="../../index_en.html">Docs</a> &raquo;</li>
      
    <li>FileManager设计文档</li>
      <li class="wy-breadcrumbs-aside">
        
          
            <a href="../../_sources/design/file_manager/README.md.txt" rel="nofollow"> View page source</a>
          
        
      </li>
  </ul>
  <hr/>
</div>
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
            
  <div class="section" id="filemanager">
<span id="filemanager"></span><h1>FileManager设计文档<a class="headerlink" href="#filemanager" title="Permalink to this headline">¶</a></h1>
<div class="section" id="">
<span id="id1"></span><h2>目标<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h2>
<p>在本文档中，我们设计说明了名为FileManager系统，方便用户上传自己的训练数据以进行分布式训练</p>
<p>主要功能包括：</p>
<ul class="simple">
<li>提供常用的命令行管理命令管理文件和目录</li>
<li>支持大文件的断点上传、下载</li>
</ul>
</div>
<div class="section" id="">
<span id="id2"></span><h2>名词解释<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li>PFS：是<code class="docutils literal"><span class="pre">Paddlepaddle</span> <span class="pre">cloud</span> <span class="pre">File</span> <span class="pre">System</span></code>的缩写，是对用户文件存储空间的抽象，与之相对的是local filesystem。目前我们用CephFS来搭建。</li>
<li><a class="reference external" href="http://docs.ceph.com/docs/master/cephfs/">CephFS</a>：一个POSIX兼容的文件系统。</li>
<li>Chunk：逻辑划上文件分块的单位。</li>
</ul>
</div>
<div class="section" id="">
<span id="id3"></span><h2>模块<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h2>
<div class="section" id="">
<span id="id4"></span><h3>架构图<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h3>
<p><image src=./src/filemanager.png width=900></p>
</div>
<div class="section" id="pfsclient">
<span id="pfsclient"></span><h3>PFSClient<a class="headerlink" href="#pfsclient" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li>功能： 详细设计<a class="reference internal" href="pfs/pfsclient.html"><span class="doc">link</span></a><ul>
<li>提供用户管理文件的命令</li>
<li>需要可以跨平台执行</li>
</ul>
</li>
<li>双向验证PFSClient需要和Ingress之间做双向验证<sup><a class="reference external" href="#tls">tls</a></sup>，所以用户需要首先在<code class="docutils literal"><span class="pre">cloud.paddlepaddle.org</span></code>上注册一下，申请用户空间，并且把系统生成的CA(certificate authority)、Key、CRT(CA signed certificate)下载到本地，然后才能使用PFSClient。</li>
</ul>
</div>
<div class="section" id="ingress">
<span id="ingress"></span><h3><a class="reference external" href="https://kubernetes.io/docs/concepts/services-networking/ingress/">Ingress</a><a class="headerlink" href="#ingress" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li>功能：提供七层协议的反向代理、基于粘性会话的负载均衡功能。</li>
<li>透传用户身份的办法Ingress需要把PFSClient的身份信息传给PFSServer，配置的方法参考<a class="reference external" href="http://www.integralist.co.uk/posts/clientcertauth.html#3">link</a></li>
</ul>
</div>
<div class="section" id="pfsserver">
<span id="pfsserver"></span><h3>PFSServer<a class="headerlink" href="#pfsserver" title="Permalink to this headline">¶</a></h3>
<p>PFSServer提供RESTful API接口，接收处理PFSClient端的文件管理请求，并且把结果返回PFSClient端。</p>
<p>RESTful API</p>
<ul class="simple">
<li>/api/v1/files<ul>
<li><code class="docutils literal"><span class="pre">GET</span> <span class="pre">/api/v1/files</span></code>: Get metadata of files or directories.</li>
<li><code class="docutils literal"><span class="pre">POST</span> <span class="pre">/api/v1/files</span></code>: Create files or directories.</li>
<li><code class="docutils literal"><span class="pre">PATCH</span> <span class="pre">/api/v1/files</span></code>: Update files or directories.</li>
<li><code class="docutils literal"><span class="pre">DELETE</span> <span class="pre">/api/v1/files</span></code>: Delete files or directories.</li>
</ul>
</li>
<li>/api/v1/file/chunks<ul>
<li><code class="docutils literal"><span class="pre">GET</span> <span class="pre">/api/v1/storage/file/chunks</span></code>: Get chunks&#8217;s metadata of a file.</li>
</ul>
</li>
<li>/api/v1/storage/files<ul>
<li><code class="docutils literal"><span class="pre">GET</span> <span class="pre">/api/v1/storage/files</span></code>: Download files or directories.</li>
<li><code class="docutils literal"><span class="pre">POST</span> <span class="pre">/api/v1/storage/files</span></code>: Upload files or directories.</li>
</ul>
</li>
<li>/api/v1/storage/file/chunks<ul>
<li><code class="docutils literal"><span class="pre">GET</span> <span class="pre">/api/v1/storage/file/chunks</span></code>: Download chunks&#8217;s data.</li>
<li><code class="docutils literal"><span class="pre">POST</span> <span class="pre">/api/v1/storage/file/chunks</span></code>: Upload chunks&#8217;s data.</li>
</ul>
</li>
</ul>
</div>
</div>
<div class="section" id="">
<span id="id5"></span><h2>文件传输优化<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h2>
<div class="section" id="">
<span id="id6"></span><h3>分块文件传输<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h3>
<p>用户文件可能是比较大的，上传到Cloud或者下载到本地的时间可能比较长，而且在传输的过程中也可能出现网络不稳定的情况。为了应对以上的问题，我们提出了Chunk的概念，一个Chunk由所在的文件偏移、数据、数据长度及校验值组成。文件的上传和下载都是通过对Chunk的操作来实现的。由于Chunk比较小（默认256K），完成一个传输动作完成的时间也比较短，不容易出错。PFSClient需要在传输完毕最后一个Chunk的时候检查destination文件的MD5值是否和source文件一致。</p>
<p>一个典型的Chunk如下所示：</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="nb">type</span> <span class="n">Chunk</span> <span class="n">struct</span> <span class="p">{</span>
    <span class="n">fileOffset</span> <span class="n">int64</span>
    <span class="n">checksum</span> <span class="n">uint32</span>
    <span class="nb">len</span>     <span class="n">uint32</span>
    <span class="n">data</span>    <span class="p">[]</span><span class="n">byte</span>
<span class="p">}</span>
</pre></div>
</div>
</div>
<div class="section" id="sparse">
<span id="sparse"></span><h3>生成sparse文件<a class="headerlink" href="#sparse" title="Permalink to this headline">¶</a></h3>
<p>当destination文件不存在或者大小和source文件不一致时，可以用<a class="reference external" href="https://Go.org/pkg/syscall/#Fallocate">Fallocate</a>生成sparse文件，然后就可以并发写入多个Chunk。</p>
</div>
<div class="section" id="">
<span id="id7"></span><h3>覆盖不一致的部分<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h3>
<p>文件传输的的关键在于需要PFSClient端对比source和destination的文件Chunks的checksum是否保持一致，不一致的由PFSClient下载或者传输Chunk完成。这样已经传输成功的部分就不用重新传输了。</p>
</div>
</div>
<div class="section" id="">
<span id="id8"></span><h2>用户使用流程<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h2>
<p>参考<a class="reference external" href="https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/cluster_train/data_dispatch.md">link</a></p>
</div>
<div class="section" id="">
<span id="id9"></span><h2>框架生成<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h2>
<p>用<a class="reference external" href="https://github.com/swagger-api/swagger-codegen">swagger</a>生成PFSClient和PFSServer的框架部分，以便我们可以把更多的精力放到逻辑本身上。</p>
</div>
<div class="section" id="">
<span id="id10"></span><h2>参考文档<a class="headerlink" href="#" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><a name=tls></a><a class="reference external" href="https://github.com/k8sp/tls/blob/master/tls.md">TLS complete guide</a></li>
<li><a class="reference external" href="http://docs.aws.amazon.com/cli/latest/reference/s3/">aws.s3</a></li>
<li><a class="reference external" href="https://linux.die.net/man/">linux man document</a></li>
</ul>
</div>
</div>


           </div>
          </div>
          <footer>
  

  <hr/>

  <div role="contentinfo">
    <p>
        &copy; Copyright 2016, PaddlePaddle developers.

    </p>
  </div>
  Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>. 

</footer>

        </div>
      </div>

    </section>

  </div>
  

    <script type="text/javascript">
        var DOCUMENTATION_OPTIONS = {
            URL_ROOT:'../../',
            VERSION:'',
            COLLAPSE_INDEX:false,
            FILE_SUFFIX:'.html',
            HAS_SOURCE:  true
        };
    </script>
      <script type="text/javascript" src="../../_static/jquery.js"></script>
      <script type="text/javascript" src="../../_static/underscore.js"></script>
      <script type="text/javascript" src="../../_static/doctools.js"></script>
      <script type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>

  
    <script type="text/javascript" src="../../_static/js/theme.js"></script>
  

  <script type="text/javascript">
      jQuery(function () {
          SphinxRtdTheme.StickyNav.enable();
      });
  </script>
   

</body>
</html>