/share/common (if you don't want
>> to modify classpaths by putting the library somewhere else)
>>
>>
>>
>> cp lzo..././target/hadoop-lzo-0.4.20-SNAPSHOT.jar ..
>> hadoop/share/hadoop/common/
>>
>>
>>
>> a. Copy /usr/lib6
.20-SNAPSHOT.jar ..
> hadoop/share/hadoop/common/
>
>
>
> a. Copy /usr/lib64/liblzo2.so.2 to .. Hadoop/lib/native/
>
>
>
>
>
> *From:* Gordon Wang [mailto:gw...@gopivotal.com]
> *Sent:* Thursday, March 06, 2014 11:50 PM
> *To:* user@hadoop.apache.org
>
mmon/
a. Copy /usr/lib64/liblzo2.so.2 to .. Hadoop/lib/native/
From: Gordon Wang [mailto:gw...@gopivotal.com]
Sent: Thursday, March 06, 2014 11:50 PM
To: user@hadoop.apache.org
Subject: Re: MR2 Job over LZO data
You can try to get the source code https://github.com/twitter/hadoop-l
You can try to get the source code
https://github.com/twitter/hadoop-lzo and then compile it against
hadoop 2.2.0.
In my memory, as long as rebuild it, lzo should work with hadoop 2.2.0
On Thu, Mar 6, 2014 at 6:29 PM, KingDavies wrote:
> Running on Hadoop 2.2.0
>
> The Java MR2 job works as ex
May be you can try download the LZO class and rebuild it against Hadoop
2.2.0;
If build success, you should be good to go;
if failed, then maybe you need to wait for the LZO guys to update their
code.
Regards,
*Stanley Shi,*
On Thu, Mar 6, 2014 at 6:29 PM, KingDavies wrote:
> Running on Hadoo
Running on Hadoop 2.2.0
The Java MR2 job works as expected on an uncompressed data source using the
TextInputFormat.class.
But when using the LZO format the job fails:
import com.hadoop.mapreduce.LzoTextInputFormat;
job.setInputFormatClass(LzoTextInputFormat.class);
Dependencies from the maven re
