Yes, you can run the collector on any host with network access to
HDFS. Verify that HADOOP_HOME and HADOOP_CONF_DIR are properly set in
chukwa-env.sh. Also verify that you can connect to hadoop from the
command line with those environment variables set.

On Tuesday, September 7, 2010, Shreya Chakravarty
<[email protected]> wrote:
> Hi,
>
> We want to setup a collector on a non HDFS machine
> Is that possible? Will the collector be able to send data to hdfs, we tried 
> setting the writer.hdfs.filesystem property to point to the hdfs master, but 
> getting the following error:
>
>
> 2010-09-07 05:53:11,153 INFO main ChukwaConfiguration - chukwaConf is 
> /home/hadoop/chukwa-0.4.0/conf/
> 2010-09-07 05:53:12,721 INFO main root - initing servletCollector
> 2010-09-07 05:53:12,727 INFO main PipelineStageWriter - using pipelined 
> writers, pipe length is 2
> 2010-09-07 05:53:12,751 INFO Thread-9 SocketTeeWriter - listen thread started
> 2010-09-07 05:53:12,755 INFO main SeqFileWriter - rotateInterval is 300000
> 2010-09-07 05:53:12,755 INFO main SeqFileWriter - outputDir is /chukwa/logs/
> 2010-09-07 05:53:12,755 INFO main SeqFileWriter - fsname is 
> iaas-1-0-16-3e-b7-d1-29:54310
> 2010-09-07 05:53:12,756 INFO main SeqFileWriter - filesystem type from 
> core-default.xml is org.apache.hadoop.hdfs.DistributedFileSystem
> 2010-09-07 05:53:12,860 ERROR main SeqFileWriter - can't connect to HDFS, 
> trying default file system instead (likely to be local)
> java.io.IOException: No FileSystem for scheme: iaas-1-0-16-3e-b7-d1-29
>         at 
> org.apache.hadoop.fs.FileSystem.createFileSystem(FileSystem.java:1375)
>         at org.apache.hadoop.fs.FileSystem.access$200(FileSystem.java:66)
>         at org.apache.hadoop.fs.FileSystem$Cache.get(FileSystem.java:1390)
>         at org.apache.hadoop.fs.FileSystem.get(FileSystem.java:196)
>         at 
> org.apache.hadoop.chukwa.datacollection.writer.SeqFileWriter.init(SeqFileWriter.java:123)
>         at 
> org.apache.hadoop.chukwa.datacollection.writer.PipelineStageWriter.init(PipelineStageWriter.java:88)
>         at 
> org.apache.hadoop.chukwa.datacollection.collector.servlet.ServletCollector.init(ServletCollector.java:112)
>         at 
> org.mortbay.jetty.servlet.ServletHolder.initServlet(ServletHolder.java:440)
>         at 
> org.mortbay.jetty.servlet.ServletHolder.doStart(ServletHolder.java:263)
>         at 
> org.mortbay.component.AbstractLifeCycle.start(AbstractLifeCycle.java:50)
>         at 
> org.mortbay.jetty.servlet.ServletHandler.initialize(ServletHandler.java:685)
>         at org.mortbay.jetty.servlet.Context.startContext(Context.java:140)
>         at 
> org.mortbay.jetty.handler.ContextHandler.doStart(ContextHandler.java:517)
>         at 
> org.mortbay.component.AbstractLifeCycle.start(AbstractLifeCycle.java:50)
>         at 
> org.mortbay.jetty.handler.HandlerWrapper.doStart(HandlerWrapper.java:130)
>         at org.mortbay.jetty.Server.doStart(Server.java:224)
>         at 
> org.mortbay.component.AbstractLifeCycle.start(AbstractLifeCycle.java:50)
>         at 
> org.apache.hadoop.chukwa.datacollection.collector.CollectorStub.main(CollectorStub.java:121)
>
> Regards
> Shreya
>
> DISCLAIMER
> ==========
> This e-mail may contain privileged and confidential information which is the 
> property of Persistent Systems Ltd. It is intended only for the use of the 
> individual or entity to which it is addressed. If you are not the intended 
> recipient, you are not authorized to read, retain, copy, print, distribute or 
> use this message. If you have received this communication in error, please 
> notify the sender and delete all copies of this message. Persistent Systems 
> Ltd. does not accept any liability for virus infected mails.
>

Reply via email to