[
https://issues.apache.org/jira/browse/NUTCH-1556?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13892423#comment-13892423
]
Koen Smets commented on NUTCH-1556:
-----------------------------------
Hi [~lewismc], I confirmed NUTCH-1679 on Cassandra store.
Although, the `bin/crawl` script changed in [NUTCH-1556] only by adding the
`$batchId` from the one in 2.2.1, this changes behaviour drastically. A lot of
pages get refetched sooner than indicated by db.default.fetch.interval and
outnumber the unfetched pages.
Although I noticed the remarkable speedup when focusing only on the pages from
the current batch, I changed `$batchId` to `-all` in order to give preference
to the pages that are truly unfetched.
> enabling updatedb to accept batchId
> ------------------------------------
>
> Key: NUTCH-1556
> URL: https://issues.apache.org/jira/browse/NUTCH-1556
> Project: Nutch
> Issue Type: Improvement
> Affects Versions: 2.2
> Reporter: kaveh minooie
> Fix For: 2.3
>
> Attachments: NUTCH-1556-batchId.patch, NUTCH-1556-v2.patch,
> NUTCH-1556-v3.patch, NUTCH-1556.patch
>
>
> So the idea here is to be able to run updatedb and fetch for different
> batchId simultaneously. I put together a patch. it seems to be working ( it
> does skip the rows that do not match the batchId), but I am worried if and
> how it might affect the sorting in the reduce part. anyway check it out.
> it also change the command line usage to this:
> Usage: DbUpdaterJob (<batchId> | -all) [-crawlId <id>]
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
