Leo,

He's using Apache Hadoop (There's a 1.0 in Apache Hadoop now:
http://hadoop.apache.org/#27+December%2C+2011%3A+release+1.0.0+available,
renamed from the earlier 0.20 series).

I assumed he was running a tarball. Great catches on the JVM parts (I
assume you looked and matched the HotSpot versions? I was too lazy to
:))!

On Wed, May 16, 2012 at 11:17 PM, Leo Leung <lle...@ddn.com> wrote:
> Hmm...
> Waqas,  I think you are using pre 1.6.0.20 build, and likely to be OpenJDK.
>
> Please try the Sun/Oracle JDK 1.6.0_26+  (as Harsh said.. don't stay away 
> from 1.7)

Typo perhaps, but to clarify, I said stay away ;-)

> And if I'm reading the logs right you are using the older release of HDP?.  
> V1.0.0
> If this is the case, you may also want to check with the HortonWorks team.
>
> Their distro datanode uses 32bit JDK and the NN uses 64bit.  You'll have to 
> be careful of running the right JVM for each node type in the same host.
>
> BTW, also I think you are starting the JT as "wagas",  try running it under 
> mapred.
> Something like
> $ su - mapred -c "<path to your hadoop-dameon.sh>  start jobtracker"  (once 
> you fix the JVM versions)
>
> Cheers
>
>
>
> -----Original Message-----
> From: Harsh J [mailto:ha...@cloudera.com]
> Sent: Wednesday, May 16, 2012 8:59 AM
> To: common-user@hadoop.apache.org
> Subject: Re: 8021 failed on connection exception: java.net.ConnectException: 
> Connection refused at 
> org.apache.hadoop.ipc.Client.wrapException(Client.java:1095)
>
> Hi,
>
> On Wed, May 16, 2012 at 9:17 PM, waqas latif <waqas...@gmail.com> wrote:
>> Hi I tried it with java 6 but with no success here are the links for
>> log and out file of jobtracker with java 6 logfile link
>> http://pastebin.com/bvWZRt0A
>>
>> outfile link is here which is a bit different from the java 7
>> http://pastebin.com/4YCZhQGh
>
> Now this does look odd (and is again the same thing). What exactly is your 
> Java6 version? i.e. java -version output?
>
> Do you also get the same error if you try a "hadoop jobtracker"
> directly in the CLI?
>
>> Also please keep in mind that i can run hadoop 0.20 with java home
>> path set to java 7.
>
> You may be able to, but none of us presently test it with that configuration. 
> So if you run into bugs or odd behavior with that, you'll pretty much be 
> alone :)
>
> --
> Harsh J



-- 
Harsh J

Reply via email to