crawlcontroller.html

用JAVA编写的,在做实验的时候留下来的,本来想删的,但是传上来,大家分享吧

<a name="276" href="#276">276</a>     <em>/**<em>*</em></em>
<a name="277" href="#277">277</a> <em>     * List of crawl status listeners.</em>
<a name="278" href="#278">278</a> <em>     *</em>
<a name="279" href="#279">279</a> <em>     * All iterations need to synchronize on this object if they're to avoid</em>
<a name="280" href="#280">280</a> <em>     * concurrent modification exceptions.</em>
<a name="281" href="#281">281</a> <em>     * See {@link java.util.Collections#synchronizedList(List)}.</em>
<a name="282" href="#282">282</a> <em>     */</em>
<a name="283" href="#283">283</a>     <strong>private</strong> <strong>transient</strong> List registeredCrawlStatusListeners =
<a name="284" href="#284">284</a>         Collections.synchronizedList(<strong>new</strong> ArrayList());
<a name="285" href="#285">285</a>     
<a name="286" href="#286">286</a>     <em class="comment">// Since there is a high probability that there will only ever by one</em>
<a name="287" href="#287">287</a>     <em class="comment">// CrawlURIDispositionListner we will use this while there is only one:</em>
<a name="288" href="#288">288</a>     <strong>private</strong> <strong>transient</strong> <a href="../../../../org/archive/crawler/event/CrawlURIDispositionListener.html">CrawlURIDispositionListener</a>
<a name="289" href="#289">289</a>         registeredCrawlURIDispositionListener;
<a name="290" href="#290">290</a> 
<a name="291" href="#291">291</a>     <em class="comment">// And then switch to the array once there is more then one.</em>
<a name="292" href="#292">292</a>      <strong>protected</strong> <strong>transient</strong> ArrayList registeredCrawlURIDispositionListeners;
<a name="293" href="#293">293</a>     
<a name="294" href="#294">294</a>     <em>/**<em>* Shared bdb Environment for <a href="../../../../org/archive/crawler/framework/Frontier.html">Frontier</a> subcomponents */</em></em>
<a name="295" href="#295">295</a>     <em class="comment">// TODO: investigate using multiple environments to split disk accesses</em>
<a name="296" href="#296">296</a>     <em class="comment">// across separate physical disks</em>
<a name="297" href="#297">297</a>     <strong>private</strong> <strong>transient</strong> Environment bdbEnvironment = <strong>null</strong>;
<a name="298" href="#298">298</a>     
<a name="299" href="#299">299</a>     <em>/**<em>*</em></em>
<a name="300" href="#300">300</a> <em>     * Shared class catalog database.  Used by the</em>
<a name="301" href="#301">301</a> <em>     * {@link #classCatalog}.</em>
<a name="302" href="#302">302</a> <em>     */</em>
<a name="303" href="#303">303</a>     <strong>private</strong> <strong>transient</strong> Database classCatalogDB = <strong>null</strong>;
<a name="304" href="#304">304</a>     
<a name="305" href="#305">305</a>     <em>/**<em>*</em></em>
<a name="306" href="#306">306</a> <em>     * Class catalog instance.</em>
<a name="307" href="#307">307</a> <em>     * Used by bdb serialization.</em>
<a name="308" href="#308">308</a> <em>     */</em>
<a name="309" href="#309">309</a>     <strong>private</strong> <strong>transient</strong> StoredClassCatalog classCatalog = <strong>null</strong>;
<a name="310" href="#310">310</a>     
<a name="311" href="#311">311</a>     <em>/**<em>*</em></em>
<a name="312" href="#312">312</a> <em>     * Keep a list of all BigMap instance made -- shouldn't be many -- so that</em>
<a name="313" href="#313">313</a> <em>     * we can checkpoint.</em>
<a name="314" href="#314">314</a> <em>     */</em>
<a name="315" href="#315">315</a>     <strong>private</strong> <strong>transient</strong> Map bigmaps = <strong>null</strong>;
<a name="316" href="#316">316</a>     
<a name="317" href="#317">317</a>     <em>/**<em>*</em></em>
<a name="318" href="#318">318</a> <em>     * Default constructor</em>
<a name="319" href="#319">319</a> <em>     */</em>
<a name="320" href="#320">320</a>     <strong>public</strong> <a href="../../../../org/archive/crawler/framework/CrawlController.html">CrawlController</a>() {
<a name="321" href="#321">321</a>         <strong>super</strong>();
<a name="322" href="#322">322</a>         <em class="comment">// Defer most setup to initialize methods</em>
<a name="323" href="#323">323</a>     }
<a name="324" href="#324">324</a> 
<a name="325" href="#325">325</a>     <em>/**<em>*</em></em>
<a name="326" href="#326">326</a> <em>     * Starting from nothing, set up CrawlController and associated</em>
<a name="327" href="#327">327</a> <em>     * classes to be ready for a first crawl.</em>
<a name="328" href="#328">328</a> <em>     *</em>
<a name="329" href="#329">329</a> <em>     * @param sH Settings handler.</em>
<a name="330" href="#330">330</a> <em>     * @throws InitializationException</em>
<a name="331" href="#331">331</a> <em>     */</em>
<a name="332" href="#332">332</a>     <strong>public</strong> <strong>void</strong> initialize(<a href="../../../../org/archive/crawler/settings/SettingsHandler.html">SettingsHandler</a> sH)
<a name="333" href="#333">333</a>     throws <a href="../../../../org/archive/crawler/framework/exceptions/InitializationException.html">InitializationException</a> {
<a name="334" href="#334">334</a>         sendCrawlStateChangeEvent(PREPARING, CrawlJob.STATUS_PREPARING);
<a name="335" href="#335">335</a> 
<a name="336" href="#336">336</a>         <strong>this</strong>.singleThreadLock = <strong>new</strong> ReentrantLock();
<a name="337" href="#337">337</a>         <strong>this</strong>.settingsHandler = sH;
<a name="338" href="#338">338</a>         <strong>this</strong>.order = settingsHandler.getOrder();
<a name="339" href="#339">339</a>         <strong>this</strong>.order.setController(<strong>this</strong>);
<a name="340" href="#340">340</a>         <strong>this</strong>.bigmaps = <strong>new</strong> Hashtable();
<a name="341" href="#341">341</a>         sExit = <span class="string">""</span>;
<a name="342" href="#342">342</a>         <strong>this</strong>.manifest = <strong>new</strong> StringBuffer();
<a name="343" href="#343">343</a>         String onFailMessage = <span class="string">""</span>;
<a name="344" href="#344">344</a>         <strong>try</strong> {
<a name="345" href="#345">345</a>             onFailMessage = <span class="string">"You must set the User-Agent and From HTTP"</span> +
<a name="346" href="#346">346</a>             <span class="string">" header values to acceptable strings. \n"</span> +
<a name="347" href="#347">347</a>             <span class="string">" User-Agent: [software-name](+[info-url])[misc]\n"</span> +
<a name="348" href="#348">348</a>             <span class="string">" From: [email-address]\n"</span>;
<a name="349" href="#349">349</a>             order.checkUserAgentAndFrom();
<a name="350" href="#350">350</a> 
<a name="351" href="#351">351</a>             onFailMessage = <span class="string">"Unable to setup disk"</span>;
<a name="352" href="#352">352</a>             <strong>if</strong> (disk == <strong>null</strong>) {
<a name="353" href="#353">353</a>                 setupDisk();
<a name="354" href="#354">354</a>             }
<a name="355" href="#355">355</a> 
<a name="356" href="#356">356</a>             onFailMessage = <span class="string">"Unable to create log file(s)"</span>;
<a name="357" href="#357">357</a>             setupLogs();
<a name="358" href="#358">358</a>             
<a name="359" href="#359">359</a>             <em class="comment">// Figure if we're to do a checkpoint restore. If so, get the</em>
<a name="360" href="#360">360</a>             <em class="comment">// checkpointRecover instance and then put into place the old bdb</em>
<a name="361" href="#361">361</a>             <em class="comment">// log files. If any of the log files already exist in target state</em>
<a name="362" href="#362">362</a>             <em class="comment">// diretory, WE DO NOT OVERWRITE (Makes for faster recovery).</em>
<a name="363" href="#363">363</a>             <em class="comment">// CrawlController checkpoint recovery code manages restoration of</em>
<a name="364" href="#364">364</a>             <em class="comment">// the old StatisticsTracker, any BigMaps used by the Crawler and</em>
<a name="365" href="#365">365</a>             <em class="comment">// the moving of bdb log files into place only. Other objects</em>
<a name="366" href="#366">366</a>             <em class="comment">// interested in recovery need to ask if</em>
<a name="367" href="#367">367</a>             <em class="comment">// CrawlController#isCheckpointRecover is set to figure if in</em>
<a name="368" href="#368">368</a>             <em class="comment">// recovery and then take appropriate recovery action</em>
<a name="369" href="#369">369</a>             <em class="comment">// (These objects can call CrawlController#getCheckpointRecover</em>
<a name="370" href="#370">370</a>             <em class="comment">// to get the directory that might hold files/objects dropped</em>
<a name="371" href="#371">371</a>             <em class="comment">// checkpointing).  Such objects will need to use a technique other</em>
<a name="372" href="#372">372</a>             <em class="comment">// than object serialization restoring settings because they'll</em>
<a name="373" href="#373">373</a>             <em class="comment">// have already been constructed when comes time for object to ask</em>
<a name="374" href="#374">374</a>             <em class="comment">// if its to recover itself. See ARCWriterProcessor for example.</em>
<a name="375" href="#375">375</a>             onFailMessage = <span class="string">"Unable to test/run checkpoint recover"</span>;
<a name="376" href="#376">376</a>             <strong>this</strong>.checkpointRecover = getCheckpointRecover();
<a name="377" href="#377">377</a>             <strong>if</strong> (<strong>this</strong>.checkpointRecover == <strong>null</strong>) {
<a name="378" href="#378">378</a>                 <strong>this</strong>.checkpointer =
<a name="379" href="#379">379</a>                     <strong>new</strong> <a href="../../../../org/archive/crawler/framework/Checkpointer.html">Checkpointer</a>(<strong>this</strong>, <strong>this</strong>.checkpointsDisk);
<a name="380" href="#380">380</a>             } <strong>else</strong> {
<a name="381" href="#381">381</a>                 setupCheckpointRecover();
<a name="382" href="#382">382</a>             }
<a name="383" href="#383">383</a>             
<a name="384" href="#384">384</a>             onFailMessage = <span class="string">"Unable to setup bdb environment."</span>;
<a name="385" href="#385">385</a>             setupBdb();
<a name="386" href="#386">386</a>             
<a name="387" href="#387">387</a>             onFailMessage = <span class="string">"Unable to setup statistics"</span>;
<a name="388" href="#388">388</a>             setupStatTracking();
<a name="389" href="#389">389</a>             
<a name="390" href="#390">390</a>             onFailMessage = <span class="string">"Unable to setup crawl modules"</span>;
<a name="391" href="#391">391</a>             setupCrawlModules();
<a name="392" href="#392">392</a>         } <strong>catch</strong> (Exception e) {
<a name="393" href="#393">393</a>             String tmp = <span class="string">"On crawl: "</span>
<a name="394" href="#394">394</a>                 + settingsHandler.getSettingsObject(<strong>null</strong>).getName() + <span class="string">" "</span> +
<a name="395" href="#395">395</a>                 onFailMessage;
<a name="396" href="#396">396</a>             LOGGER.log(Level.SEVERE, tmp, e);
<a name="397" href="#397">397</a>             <strong>throw</strong> <strong>new</strong> <a href="../../../../org/archive/crawler/framework/exceptions/InitializationException.html">InitializationException</a>(tmp, e);
<a name="398" href="#398">398</a>         }
<a name="399" href="#399">399</a> 
<a name="400" href="#400">400</a>         <em class="comment">// force creation of DNS Cache now -- avoids CacheCleaner in toe-threads group</em>
<a name="401" href="#401">401</a>         dns.getRecords(<span class="string">"localhost"</span>, Type.A, DClass.IN);
<a name="402" href="#402">402</a>         
<a name="403" href="#403">403</a>         setupToePool();
<a name="404" href="#404">404</a>         setThresholds();
<a name="405" href="#405">405</a>         
<a name="406" href="#406">406</a>         reserveMemory = <strong>new</strong> LinkedList();
<a name="407" href="#407">407</a>         <strong>for</strong>(<strong>int</strong> i = 1; i &lt; RESERVE_BLOCKS; i++) {
<a name="408" href="#408">408</a>             reserveMemory.add(<strong>new</strong> <strong>char</strong>[RESERVE_BLOCK_SIZE]);
<a name="409" href="#409">409</a>         }
<a name="410" href="#410">410</a>     }
<a name="411" href="#411">411</a>     
<a name="412" href="#412">412</a>     <em>/**<em>*</em></em>
<a name="413" href="#413">413</a> <em>     * Does setup of checkpoint recover.</em>
<a name="414" href="#414">414</a> <em>     * Copies bdb log files into state dir.</em>
<a name="415" href="#415">415</a> <em>     * @throws IOException</em>
<a name="416" href="#416">416</a> <em>     */</em>
<a name="417" href="#417">417</a>     <strong>protected</strong> <strong>void</strong> setupCheckpointRecover()