NoSQL服务器 Cassandra 1.0 稳定版发布

jopen 13年前
     <p>Apache软件基金会(ASF)发布了 NoSQL 数据库 Cassandra 1.0 正式版。</p>    <p><strong>主要新特性包括:</strong></p>    <p>1. Windows 下的服务安装,只需运行:bin\cassandra.bat install<br /> 2. Hinted handoff<br /> 3. 多线程压缩<br /> 4. 节点替换<br /> 5. 全新的 CQL 客户端</p>    <p>详细的改进内容请看<span style="text-decoration:underline;"><br /> </span></p>    <p>Besides <a href="/misc/goto?guid=4958194562911439568">compression</a>, <a href="/misc/goto?guid=4958194563675091001">memory and disk space management</a>, <a href="/misc/goto?guid=4958194564427775209">leveled compaction</a>, and <a href="/misc/goto?guid=4958194565168969838">performance</a> improvements, Cassandra 1.0 also includes a number of smaller features and enhancements.</p>    <h3>Windows service</h3>    <p>You can still run Cassandra 1.0 from the bat file if you want — this is the most convenient way to develop and run quick tests — but for production, installing it as a Windows service is a much better option. To do this, just run <tt>bin\cassandra.bat install</tt>. (Similarly, uninstall will remove it.)</p>    <h3>Hinted handoff</h3>    <p><a href="/misc/goto?guid=4958194565911794687">Hinted handoff</a> is one way Cassandra replays missed updates to nodes that were temporarily down or unreachable. Cassandra 1.0 makes two improvements:</p>    <ul>     <li>Hint data is <a href="/misc/goto?guid=4958194566646852656">stored more efficiently</a>, reducing the load imposed by hint replay on nodes storing them</li>     <li><a href="/misc/goto?guid=4958194567381802586">Coordinator</a> nodes no longer need to wait for the <a href="/misc/goto?guid=4958194568123172960">failure detector</a> to recognize a problem with a replication target, to begin generating hints for it</li>    </ul>    <p>This means that <a href="/misc/goto?guid=4958194568858001604">full repair</a>, while still recommended periodically, will only have to reconcile data missing because of a coordinator failure at the same time, or because of losing a node entirely. Both of these are relatively rare occurrences. </p>    <p>Similarly, <a href="/misc/goto?guid=4958194569595333777">read repair</a> is less necessary, so the default read repair probability has been reduced to 10% from 100%. This will result in significantly improved throughput in clusters with more than one replica and lower <a href="/misc/goto?guid=4958194570333947608"><tt>ConsistencyLevel</tt></a> reads than ALL.</p>    <h3>Multithreaded compaction</h3>    <p>Cassandra 0.8 introduced concurrent <a href="/misc/goto?guid=4958194571072804250">compaction</a>: running multiple compaction sets at a time, each in its own thread. Cassandra 1.0 adds multithreaded compaction, which allows multiple threads to be used per compaction set.</p>    <p>The primary use case for multithreaded compaction is when you either have a small number of column families or you are using <a href="/misc/goto?guid=4958194564427775209">leveled compaction</a>, and you want to keep your data as compacted as possible--usually because you have a read-heavy workload.</p>    <p>Multithreaded compaction can be enabled by setting <tt>multithreaded_compaction</tt> to true in <tt>cassandra.yaml</tt>.</p>    <h3>Node replacement</h3>    <p>Replacing a node was a <a href="/misc/goto?guid=4958194572495899376">two-step process</a> in earlier versions of Cassandra: one to add the new node, and a second to remove the old. Cassandra 1.0 simplifies this common operation into simply starting the new node with the option <tt>-Dcassandra.replace_token=[token]</tt>.</p>    <h3>New CQL clients</h3>    <p>After some <a href="/misc/goto?guid=4958194573227317314">discussion</a>, we decided that <a href="/misc/goto?guid=4958194573971143621">CQL</a> drivers would be better developed as separate projects rather than in the Cassandra tree. The <a href="/misc/goto?guid=4958194574709001116">Java</a> and <a href="/misc/goto?guid=4958194575453079768">Python</a> drivers have moved to <a href="/misc/goto?guid=4958194576189242146">Apache Extras</a> on code.google.com, and new <a href="/misc/goto?guid=4958194576933344517">Ruby</a> and <a href="/misc/goto?guid=4958194577673364741">PHP</a> drivers have joined them, with <a href="/misc/goto?guid=4958194578408758548">node.js</a> coming soon.</p>    <h3>Previously</h3>    <ul>     <li><a href="/misc/goto?guid=4958194565168969838">What's new in Cassandra 1.0, part 4: Performance</a> </li>     <li><a href="/misc/goto?guid=4958194564427775209">What's new in Cassandra 1.0, part 3: Leveled compaction</a> </li>     <li><a href="/misc/goto?guid=4958194563675091001">What's new in Cassandra 1.0, part 2: Improved memory and disk space management</a> </li>     <li><a href="/misc/goto?guid=4958194562911439568">What's new in Cassandra 1.0, part 1: Compression</a> </li>     <li><a href="/misc/goto?guid=4958194581853983722">What's new in Cassandra 0.8</a> </li>     <li><a href="/misc/goto?guid=4958194582601144608">What's new in Cassandra 0.7</a> </li>     <li><a href="/misc/goto?guid=4958194583330266059">What's new in Cassandra 0.6</a> </li>    </ul>    <p> </p>    <p><strong>下载地址:</strong><a href="/misc/goto?guid=4958194584086650239" target="_blank">Apache Cassandra v1.0</a></p>    <div id="p_fullcontent" class="detail">     <p><strong><br /> <img title="Apache Cassandra是一套开源分布式Key-Value存储系统。" border="0" alt="Apache Cassandra是一套开源分布式Key-Value存储系统。" src="https://simg.open-open.com/show/436d8cfdac996a37ddd10929b6de4959.png" width="500" height="100" /><br /> Apache Cassandra</strong>是一套<strong>开源分布式Key-Value存储系统</strong>。它最初由非死book开发,用于储存特别大的数据。<strong>非死book目前在使用此系统</strong>。</p>     <p>主要特性:</p>     <ul>      <li>分布式</li>      <li>基于column的结构化</li>      <li>高伸展性</li>     </ul>     <p>Cassandra的主要特点就是它不是一个数据库,而是由一堆数据库节点共同构成的一个分布式网络服务,对Cassandra 的一个写操作,会被复制到其他节点上去,对Cassandra的读操作,也会被路由到某个节点上面去读取。对于一个Cassandra群集来说,扩展性能 是比较简单的事情,只管在群集里面添加节点就可以了。</p>     <p>Cassandra是一个混合型的非关系的数据库,类似于Google的BigTable。其主要功能比 Dynomite(分布式的Key-Value存 储系统)更丰富,但支持度却不如文档存储MongoDB(介于关系数据库和非关系数据库之间的开源产品,是非关系数据库当中功能最丰富,最像关系数据库 的。支持的数据结构非常松散,是类似json的bjson格式,因此可以存储比较复杂的数据类型。)Cassandra最初由非死book开发,后转变成了开源项目。它是一个网络社交云计算方面理想的数据库。以Amazon专有的完全分布式的Dynamo为基础,结合了Google BigTable基于列族(Column Family)的数据模型。P2P去中心化的存储。很多方面都可以称之为Dynamo 2.0。</p>     <p>和其他数据库比较,有几个突出特点:</p>     <p><strong>模式灵活</strong> :使用Cassandra,像文档存储,你不必提前解决记录中的字段。你可以在系统运行时随意的添加或移除字段。这是一个惊人的效率提升,特别是在大型部 署上。 <br /> <strong>真正的可扩展性</strong> :Cassandra是纯粹意义上的水平扩展。为给集群添加更多容量,可以指向另一台电脑。你不必重启任何进程,改变应用查询,或手动迁移任何数据。 <br /> <strong>多数据中心识别</strong> :你可以调整你的节点布局来避免某一个数据中心起火,一个备用的数据中心将至少有每条记录的完全复制。</p>     <p>一些使Cassandra提高竞争力的其他功能:</p>     <p><strong>范围查询</strong> :如果你不喜欢全部的键值查询,则可以设置键的范围来查询。 <br /> <strong>列表数据结构</strong> :在混合模式可以将超级列添加到5维。对于每个用户的索引,这是非常方便的。 <br /> <strong>分布式写操作</strong> :有可以在任何地方任何时间集中读或写任何数据。并且不会有任何单点失败。<br /> <br /> <span style="font-weight:bold;">项目地址:</span><a style="font-weight:bold;" href="/misc/goto?guid=4958189053392575275" target="_blank">http://cassandra.apache.org/</a></p>    </div>