Protocol plugins should report progress to the fetcher
------------------------------------------------------

                 Key: NUTCH-569
                 URL: https://issues.apache.org/jira/browse/NUTCH-569
             Project: Nutch
          Issue Type: Improvement
          Components: fetcher
    Affects Versions: 1.0.0
            Reporter: Andrzej Bialecki 
            Priority: Minor


When downloading very large files over slow connections, protocol plugins spend 
long time in Protocol.getProtocolOutput(...). This sometimes leads to a timeout 
in Fetcher / Fetcher2, with the message "aborting with hung threads". Protocol 
plugins should periodically notify their caller about progress. In a situation 
when the call to getProtocolOutput takes very long time to return, this will 
help the caller to determine whether the wait is justified.

Preferably, the callback interface should allow the monitoring of not only the 
binary progress / no-progress, but also the download speed, so that the caller 
could terminate slow connections. E.g.

{noformat}
interface ProtocolReporter {
  void progress(long bytesDownloaded);
}
{noformat}


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

Reply via email to