[
https://issues.apache.org/jira/browse/NUTCH-2084?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14710828#comment-14710828
]
Markus Jelsma commented on NUTCH-2084:
--------------------------------------
Well, this immediately helped me track down bad segments. Will commit shortly.
> Track changes in input dirs for SegmentMerger
> ---------------------------------------------
>
> Key: NUTCH-2084
> URL: https://issues.apache.org/jira/browse/NUTCH-2084
> Project: Nutch
> Issue Type: Bug
> Affects Versions: 1.10
> Reporter: Markus Jelsma
> Fix For: 1.11
>
> Attachments: NUTCH-2084.patch
>
>
> When merging 1000's of segments, and one is corrupt, broken, whatever, the
> merge goes with the least number of shared input dirs, or nothing. This patch
> tracks when input dirs change due to a segment missing any directory.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
