Sure Zhang.
Thanks for help.
-Rahul
On Aug 26, 2010, at 8:17 PM, Jeff Zhang wrote:
> It's weird. I doubt maybe there's other configuration file on your
> class path which override your real conf files.
> Could you download a new pig release and follow the instructions on
> http://hadoop.apache
It's weird. I doubt maybe there's other configuration file on your
class path which override your real conf files.
Could you download a new pig release and follow the instructions on
http://hadoop.apache.org/pig/docs/r0.7.0/setup.html on a new
environment ?
On Thu, Aug 26, 2010 at 7:49 PM, rahu
Hi ,
I tried the grunt shell as well but that also does not connects to hadoop. It
throws a warning and runs the job in standalone mode. So it tried it using the
pig.jar.
Do you have any further suggestion on that ?
Rahul
On Aug 26, 2010, at 7:23 PM, Jeff Zhang wrote:
> Connect to 9001 is ri
Connect to 9001 is right, this is jobtracker's ipc port while 50030
is its http server port.
And have you ever try to run the grunt shell ?
On Thu, Aug 26, 2010 at 7:12 PM, rahul wrote:
> Hi Jeff,
>
> I can connect to the jobtracker web UI using the following URL :
> http://localhost:50030/jobt
Hi Jeff,
I can connect to the jobtracker web UI using the following URL :
http://localhost:50030/jobtracker.jsp
And also I can see jobs which I ran directly using the streaming api on hadoop.
I also see it tries to connect to localhost/127.0.0.1:9001 which I have
specified in the hadoop conf f
Can you look at the jobtracker log or access jobtracker web ui ?
It seems you can not connect to jobtracker according your log
"Caused by: java.io.IOException: Call to localhost/127.0.0.1:9001
failed on local exception: java.io.EOFException"
On Fri, Aug 27, 2010 at 10:00 AM, rahul wrote:
> Ye
Yes they are running.
On Aug 26, 2010, at 6:59 PM, Jeff Zhang wrote:
> Execute command jps in shell to see whether namenode and jobtracker is
> running correctly.
>
>
>
> On Fri, Aug 27, 2010 at 9:49 AM, rahul wrote:
>> Hi Jeff,
>>
>> I transferred the hadoop conf files to the pig/conf locat
Execute command jps in shell to see whether namenode and jobtracker is
running correctly.
On Fri, Aug 27, 2010 at 9:49 AM, rahul wrote:
> Hi Jeff,
>
> I transferred the hadoop conf files to the pig/conf location but still i get
> the same error.
>
> Does the issue is with the configuration fil
wrote:
> Can you try replacing localhost with the fully qualified name of your host?
>
> Santhosh
>
>
> -Original Message-
> From: rahul [mailto:rmalv...@apple.com]
> Sent: Thursday, August 26, 2010 6:22 PM
> To: pig-user@hadoop.apache.org
> Subject: Re: Pi
Hi Jeff,
I transferred the hadoop conf files to the pig/conf location but still i get
the same error.
Does the issue is with the configuration files or with the hdfs files system ?
Can test the connection to hdfs(localhost/127.0.0.1:9001) in some way ?
Steps I did :
1. I have formatted init
Try to put the hadoop xml configuration file to pig/conf folder
On Thu, Aug 26, 2010 at 6:22 PM, rahul wrote:
> Hi Jeff,
>
> I have set the hadoop conf in class path by setting $HADOOP_CONF_DIR variable.
>
> But I have both Pig and hadoop running at the same machine, so localhost
> should not
But according the errer log:
"Could not validate the output specification for:
file:///Users/rahulmalviya/Documents/Pig/dev/main_merged_hdp_out"
It still try to access local file system rather than HDFS File System
On Thu, Aug 26, 2010 at 6:22 PM, rahul wrote:
> Hi Jeff,
>
> I have set the had
Can you try replacing localhost with the fully qualified name of your host?
Santhosh
-Original Message-
From: rahul [mailto:rmalv...@apple.com]
Sent: Thursday, August 26, 2010 6:22 PM
To: pig-user@hadoop.apache.org
Subject: Re: Pig and Hadoop Integration Error
Hi Jeff,
I have set
Hi Jeff,
I have set the hadoop conf in class path by setting $HADOOP_CONF_DIR variable.
But I have both Pig and hadoop running at the same machine, so localhost should
not make a difference.
So I have used all the default config setting for the core-site.xml,
hdfs-site.xml, mapred-site.xml, as
Do you put the hadoop conf on classpath ? It seems you are still using
local file system but conncect Hadoop's JobTracker.
Make sure you set the correct configuration in core-site.xml
hdfs-site.xml, mapred-site.xml, and put them on classpath.
On Thu, Aug 26, 2010 at 5:32 PM, rahul wrote:
> Hi ,
Hi ,
I am trying to integrate Pig with Hadoop for processing of jobs.
I am able to run Pig in local mode and Hadoop with streaming api perfectly.
But when I try to run Pig with Hadoop I get follwong Error:
Pig Stack Trace
---
ERROR 2116: Unexpected error. Could not validate the outp
16 matches
Mail list logo
