[jira] Resolved: (CONNECTORS-28) When RSS connector fails to fetch a feed due to network difficulties, its retry strategy is flawed

["Karl Wright (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/CONNECTORS-28?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Karl Wright resolved CONNECTORS-28.
-----------------------------------

    Resolution: Fixed

Fix committed in r939308.  This is a narrow fix designed to retry only on NoRouteToHostException, which is the exact case I was seeing.  Other generic IOExceptions still have the stated behavior, because that prevents blind retries when we don't understand what happened.


> When RSS connector fails to fetch a feed due to network difficulties, its retry strategy is flawed
> --------------------------------------------------------------------------------------------------
>
>                 Key: CONNECTORS-28
>                 URL: https://issues.apache.org/jira/browse/CONNECTORS-28
>             Project: Lucene Connector Framework
>          Issue Type: Bug
>          Components: RSS connector
>            Reporter: Karl Wright
>            Assignee: Karl Wright
>
> For a continuous RSS crawl, when the RSS connector cannot fetch a feed due to a socket connection problem (e.g. error -103), it does not schedule retry of the feed fetch in a feed-specific way, but rather uses the default document refetch interval.  So, if someone has configured their job to never refetch documents, the feed effectively is never tried again.
> The behavior should be for the fetch to be retried for some period of time before giving up.  Since the connector does not know what type of document it is fetching in advance, all document fetches must be treated in this way in this situation.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.