hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Philip Zeyliger <phi...@cloudera.com>
Subject Re: child task choking on large (19K) input line
Date Wed, 17 Jun 2009 18:43:37 GMT
On Fri, Jun 5, 2009 at 1:41 PM, Thushara Wijeratna<thushw@gmail.com> wrote:
> i fixed this on my system - Text.append degraded for large buffers (225M
> line could not be processed for over 20 min).
> the fix was to substitute a ByteArrayOutputStream and now the map/reduce
> runs in under 30s.
> should i submit a patch for this? what is the process?

Please do contribute patches!  The instructions are at
http://wiki.apache.org/hadoop/HowToContribute .  You'll end up
submitting a JIRA, writing a patch, including some tests (and, for
performance issues, some performance numbers and steps to reproduce).

-- Philip

View raw message