It do see stuff about crc's being ignored sometimes at the end of
an operation. Is there a setting for this?
Isn't it in the nutch-default.xml?
I have also just learned that the box i've been using as the job
tracker and ndfs name node has a wonky system timer. So maybe that
is the problem. I'm currently getting a test running using only the
other two machines.
Well this could may be the problem, I'm not sure how java handles the
system current millis behind the sense. Since you got a time out
exception may this could be a problem.
Stefan
-Matt
Stefan Groschupf wrote:
hmm, sounds strange, but I'm interested to dig to find the
problem source, since I'm very much interested to get 0.8 stabile
asap.
However find such a problem source is a pain in the neck.
Do you use the latest sources from svn?
Do you ignore crc errors? Doug mentioned that he notice often
problems with this.
Stefan
Am 13.12.2005 um 20:28 schrieb Matt Zytaruk:
I dont think the network settings are the problem, as I have
been able to parse other segments using map reduce no problem.
If it was the network configuration, wouldn't it never work?
However, things do not seem to be stable, as some operations in
ndfs will error, and then I do the same thing 5 minutes later
and it works fine. Same with other things, some crawls work
fine, others throw exceptions and crash (I actually had a crawl
crash with the same problem as below). This is using 3 Opteron
boxes running Suse Linux.
-Matt Zytaruk
Stefan Groschupf wrote:
Looks like a problem with the tcp ip communication.
Any firewalls running on the boxes? May any ports closed?
Are the dns names correct configured?
Is your job tracker running stable?
Stefan
Am 13.12.2005 um 19:46 schrieb Matt Zytaruk:
Hello all, I've been trying to parse a segment of data
(probably around 500k pages) I previously fetched, and
everytime I try, I get an error. Below is the error given by
the slaves. The master gives a similar error. This usually
happens late in the reduce phase, but has also happened
during the map phase once. Any ideas what might be going on
here? Network issues? bugs in the tracker?
Thanks for any help you might be able to give.
-matt zytaruk
Slaves:
060102 200647 task_m_bvkze5 Child Error
java.io.IOException: Task process exit with nonzero status.
at org.apache.nutch.mapred.TaskRunner.runChild
(TaskRunner.java:139)
at org.apache.nutch.mapred.TaskRunner.run
(TaskRunner.java:92)
060102 200833 task_m_bvkze5 done; removing files.
060102 200855 Client connection to 64.141.15.126:8050: closing
java.lang.reflect.UndeclaredThrowableException
at $Proxy0.pollForClosedTask(Unknown Source)
at org.apache.nutch.mapred.TaskTracker.offerService
(TaskTracker.java:241)
at org.apache.nutch.mapred.TaskTracker.run
(TaskTracker.java: 268)
at org.apache.nutch.mapred.TaskTracker.main
(TaskTracker.java: 633)
Caused by: java.io.IOException: timed out waiting for response
at org.apache.nutch.ipc.Client.call(Client.java:296)
at org.apache.nutch.ipc.RPC$Invoker.invoke(RPC.java:127)
... 4 more
060102 201229 Lost connection to JobTracker [crawler-
d-01.internal.wavefire.ca/64.141.15.126:8050]. Retrying...
Master:
Exception in thread "main"
java.lang.reflect.UndeclaredThrowableException
at $Proxy0.getJobStatus(Unknown Source)
at org.apache.nutch.mapred.JobClient.getJob
(JobClient.java: 272)
at org.apache.nutch.mapred.JobClient.runJob
(JobClient.java: 295)
at org.apache.nutch.crawl.ParseSegment.parse
(ParseSegment.java:91)
at org.apache.nutch.crawl.ParseSegment.main
(ParseSegment.java:110)
Caused by: java.io.IOException: timed out waiting for response
at org.apache.nutch.ipc.Client.call(Client.java:296)
at org.apache.nutch.ipc.RPC$Invoker.invoke(RPC.java:127)
---------------------------------------------------------------
company: http://www.media-style.com
forum: http://www.text-mining.org
blog: http://www.find23.net
---------------------------------------------------------------
company: http://www.media-style.com
forum: http://www.text-mining.org
blog: http://www.find23.net
