Hi, Ed,

Funny you should choose just now to reply. I just solved the problem on my
own system and was about to post what I found.

This appears to be related to HADOOP-406:
https://issues.apache.org/jira/browse/HADOOP-406

That appears to be why the child JVM fails to inherit hadoop.log.dir, etc.

I applied the patch for that, and then added the following to
mapred.child.java.opts in  conf/hadoop-site.xml:
[EMAIL PROTECTED]@
[EMAIL PROTECTED]@

I also made my HADOOP_LOG_DIR environment var an absolute path instead of a
relative one, so that any inheriting processes don't care where they're
running (that did seem to be a problem, too).

I don't know if Chris Schneider's patch for HADOOP-406 will prove to be the
"long-term" solution, but it certainly works for me.

Thanks for following up!

Doug


e w wrote:
> 
> Hi Doug,
> 
> There was a discussion under the subject "log4j.properties bug (?)" a
> couple
> of weeks back. Please check it out. My (temporary) solution was to
> hardwire
> the "log4j.appender.DRFA.File" variable in log4j.properties to
> hadoop.logand then all the fetcher output from all tasks gets written
> there as well.
> Not an elegant solution but it works for debugging purposes. The problem
> seems to be to do with the environment variable ${hadoop.log.dir} not
> being
> set when log4j.properties is parsed and so nutch tries to write to a file
> in
> "/".
> 
> -Ed
> 
> On 8/22/06, Doug Cook <[EMAIL PROTECTED]> wrote:
>>
>>
>> Hi, Ed-
>>
>> I'm seeing the same problem. If anyone has had a similar experience and
>> solved it, please let me know. In the mean time, I'll keep investigating
>> and
>> post back if I figure out what's going wrong.
>>
>> This may or may not matter, but I'm running everything on a single MP
>> machine w/o DFS.
>>
>> Doug
>>
>>
>> e w wrote:
>> >
>> > Logging of the Fetcher output in 0.8-dev used to work (writing to the
>> > corresponding tasktracker output log) but doesn't appear to any more
>> with
>> > the nightly build from a couple of weeks ago and also the one from last
>> > night.
>> >
>> > I've enabled DEBUG for the first 4 logging properties in
>> > conf/log4j.properties and also:
>> >
>> > <property>
>> >   <name>fetcher.verbose</name>
>> >   <value>true</value>
>> >   <description>If true, fetcher will log more verbosely.</description>
>> > </property>
>> >
>> > in nutch-site.xml.
>> >
>> > Does anyone else have this problem? If not, can anyone suggest a likely
>> > cause? Thanks,
>> >
>> > -Ed
>> >
>> >
>>
>> --
>> View this message in context:
>> http://www.nabble.com/Problem-with-logging-of-Fetcher-output-in-0.8-dev-tf1999249.html#a5914756
>> Sent from the Nutch - User forum at Nabble.com.
>>
>>
> 
> 

-- 
View this message in context: 
http://www.nabble.com/Problem-with-logging-of-Fetcher-output-in-0.8-dev-tf1999249.html#a5956385
Sent from the Nutch - User forum at Nabble.com.


-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Nutch-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/nutch-general

Reply via email to