NullPointerException at "NewTrackingRecordReader.initialize(MapTask.java:548)"

[TheWall89]

Hi, I'm trying to use hadoop-pcap to work with some pcap files on hadoop, but I get the issue in the title.

Here follows my configuration.

I'm using cloudera 5.5 virtual machine and I compiled hadoop-pcap by cloning the repo and issuing the following command:
mvn -Dskip-tests clean package
This produced the file hadoop-pcap-lib-1.2-SNAPSHOT.jar.

I imported this file in Eclipse and I rewrote your example found here to respect new hadoop standards. Here is my code (I had to attach .txt files because github doesn't accept .java)
PcapCount.txt
PcapMapper.txt
PcapReducer.txt

With Eclipse I created a jar file (pcapcount.jar) and I tried to execute it in hadoop with the following commands:

export HADOOP_CLASSPATH=/home/cloudera/pcap-tests/hadoop-pcap-lib-1.2-SNAPSHOT.jar
hadoop jar /home/cloudera/pcap-tests/pcapcount.jar mypkg.PcapCount -libjars /home/cloudera/pcap-tests/hadoop-pcap-lib-1.2-SNAPSHOT.jar /user/cloudera/pcapcount/input /user/cloudera/pcapcount/output

The mapreduce job can't start and this is the error trace: ErrorTrace.txt.

I searched for the error in the web and I found this useful information:

1. This should be the class where the NullPointerException happens (line 548): org.apache.hadoop.mapred.MapTask
2. I think the problem is in your class PcapInputFormat.java. As stated in this thread, Hadoop: NullPointerException with Custom InputFormat, your custom InputFormat should set the backend variable.

It would be great if you can look into this or give me some advice. I am quite new to hadoop.
Thanks.