爬虫突然死亡时如何保持状态？

Question

这个公式是引用抓取蜘蛛不存储状态(持久状态)的

我按照下面的链接来保持爬虫http://doc.scrapy.org/en/latest/topics/jobs.html的状态

现在，当爬行器在中断或Ctrl+C中正确结束时，这是非常好的工作方式。

我注意到蜘蛛没有在

1. 你击中了Ctrl +C多次。
2. 服务器容量被击中。
3. 它突然结束的任何其他原因

当蜘蛛再次运行时，它会在爬起来的第一个url上关闭自己。

当发生上述情况时，如何实现爬行器的持久状态？否则它就会再次爬上一整堆urls。

当蜘蛛再次运行时记录：

2016-08-30 08:14:11 [scrapy] INFO: Scrapy 1.1.2 started (bot: maxverstappen)
2016-08-30 08:14:11 [scrapy] INFO: Overridden settings: {'NEWSPIDER_MODULE': 'maxverstappen.spiders', 'SPIDER_MODULES': ['maxverstappen.spiders'], 'ROBOTSTXT_OBEY': True, 'BOT_NAME': 'maxverstappen'}
2016-08-30 08:14:11 [scrapy] INFO: Enabled extensions:
['scrapy.extensions.logstats.LogStats',
 'scrapy.extensions.telnet.TelnetConsole',
 'scrapy.extensions.corestats.CoreStats',
 'scrapy.extensions.spiderstate.SpiderState']
2016-08-30 08:14:11 [scrapy] INFO: Enabled downloader middlewares:
['scrapy.downloadermiddlewares.robotstxt.RobotsTxtMiddleware',
 'scrapy.downloadermiddlewares.httpauth.HttpAuthMiddleware',
 'scrapy.downloadermiddlewares.downloadtimeout.DownloadTimeoutMiddleware',
 'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware',
 'scrapy.downloadermiddlewares.retry.RetryMiddleware',
 'scrapy.downloadermiddlewares.defaultheaders.DefaultHeadersMiddleware',
 'scrapy.downloadermiddlewares.redirect.MetaRefreshMiddleware',
 'scrapy.downloadermiddlewares.httpcompression.HttpCompressionMiddleware',
 'scrapy.downloadermiddlewares.redirect.RedirectMiddleware',
 'scrapy.downloadermiddlewares.cookies.CookiesMiddleware',
 'scrapy.downloadermiddlewares.chunked.ChunkedTransferMiddleware',
 'scrapy.downloadermiddlewares.stats.DownloaderStats']
2016-08-30 08:14:11 [scrapy] INFO: Enabled spider middlewares:
['scrapy.spidermiddlewares.httperror.HttpErrorMiddleware',
 'scrapy.spidermiddlewares.offsite.OffsiteMiddleware',
 'scrapy.spidermiddlewares.referer.RefererMiddleware',
 'scrapy.spidermiddlewares.urllength.UrlLengthMiddleware',
 'scrapy.spidermiddlewares.depth.DepthMiddleware']
2016-08-30 08:14:12 [scrapy] INFO: Enabled item pipelines:
['maxverstappen.pipelines.MaxverstappenPipeline']
2016-08-30 08:14:12 [scrapy] INFO: Spider opened
2016-08-30 08:14:12 [scrapy] INFO: Crawled 0 pages (at 0 pages/min), scraped 0 items (at 0 items/min)
2016-08-30 08:14:12 [scrapy] DEBUG: Telnet console listening on 127.0.0.1:6024
2016-08-30 08:14:12 [scrapy] DEBUG: Crawled (200) <GET http://www.inautonews.com/robots.txt> (referer: None)
2016-08-30 08:14:12 [scrapy] DEBUG: Crawled (200) <GET http://www.thecheckeredflag.com/robots.txt> (referer: None)
2016-08-30 08:14:12 [scrapy] DEBUG: Crawled (200) <GET http://www.inautonews.com/> (referer: None)
2016-08-30 08:14:12 [scrapy] DEBUG: Crawled (200) <GET http://www.thecheckeredflag.com/> (referer: None)
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered duplicate request: <GET http://www.inautonews.com/> - no more duplicates will be shown (see DUPEFILTER_DEBUG to show all duplicates)
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.newsnow.co.uk': <GET http://www.newsnow.co.uk/h/Life+&+Style/Motoring>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.americanmuscle.com': <GET http://www.americanmuscle.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.extremeterrain.com': <GET http://www.extremeterrain.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.autoanything.com': <GET http://www.autoanything.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.bmwcoop.com': <GET http://www.bmwcoop.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.automotorblog.com': <GET http://www.automotorblog.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'twitter.com': <GET https://twitter.com/inautonews>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.facebook.com': <GET https://www.facebook.com/inautonews>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'plus.google.com': <GET https://plus.google.com/+Inautonewsplus>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.histats.com': <GET http://www.histats.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.hamiltonf1site.com': <GET http://www.hamiltonf1site.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.joshwellsracing.com': <GET http://www.joshwellsracing.com/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.jensonbuttonfan.net': <GET http://www.jensonbuttonfan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.fernandoalonsofan.net': <GET http://www.fernandoalonsofan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.markwebberfan.net': <GET http://www.markwebberfan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.felipemassafan.net': <GET http://www.felipemassafan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.nicorosbergfan.net': <GET http://www.nicorosbergfan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.nickheidfeldfan.net': <GET http://www.nickheidfeldfan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.lewishamiltonblog.net': <GET http://www.lewishamiltonblog.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.timoglockfan.net': <GET http://www.timoglockfan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.jarnotrullifan.net': <GET http://www.jarnotrullifan.net/>
2016-08-30 08:14:12 [scrapy] DEBUG: Filtered offsite request to 'www.brunosennafan.net': <GET http://www.brunosennafan.net/>
2016-08-30 08:14:12 [scrapy] INFO: Closing spider (finished)
2016-08-30 08:14:12 [scrapy] INFO: Dumping Scrapy stats:
{'downloader/request_bytes': 896,
 'downloader/request_count': 4,
 'downloader/request_method_count/GET': 4,
 'downloader/response_bytes': 35353,
 'downloader/response_count': 4,
 'downloader/response_status_count/200': 4,
 'dupefilter/filtered': 149,
 'finish_reason': 'finished',
 'finish_time': datetime.datetime(2016, 8, 30, 8, 14, 12, 724932),
 'log_count/DEBUG': 28,
 'log_count/INFO': 7,
 'offsite/domains': 22,
 'offsite/filtered': 23,
 'request_depth_max': 1,
 'response_received_count': 4,
 'scheduler/dequeued': 2,
 'scheduler/dequeued/disk': 2,
 'scheduler/enqueued': 2,
 'scheduler/enqueued/disk': 2,
 'start_time': datetime.datetime(2016, 8, 30, 8, 14, 12, 13456)}
2016-08-30 08:14:12 [scrapy] INFO: Spider closed (finished)

Answer 1

这样做的一种方法是通过有两个蜘蛛来区分发现和消费者逻辑。一种是发现产品urls，另一种是使用这些urls，并返回每个urls的结果。如果由于某种原因，使用者在运行过程中死亡，它可以很容易地恢复爬行，因为发现队列不受此崩溃的影响。

已经有一个伟大的工具，刮刮的家伙，正是这样做。它叫前翅目

Frontera是一个由爬行前沿和分布/缩放基元组成的web爬行框架，允许构建大规模的在线web爬虫。 Frontera负责在爬行过程中遵循的逻辑和策略。它存储爬虫提取的链接并对其排序，以决定下一步访问哪些页面，并能够以分布式方式进行访问。

这听起来很复杂，但却相当直截了当。但是，如果您运行了一些小规模的，并且关闭了一个，您可能只想手动处理这个问题。您可以运行发现蜘蛛并在json中输出结果，然后在使用者蜘蛛中以持久的方式解析json (即从它中弹出值)。