Skip to content

Latest commit

 

History

History
35 lines (24 loc) · 999 Bytes

README.md

File metadata and controls

35 lines (24 loc) · 999 Bytes
Raw

hadoopifications

A small hadoop job to create CDX files for NetarchiveSuite

hadoop jar hadoopifications-1.0-SNAPSHOT-hadoop.jar \
    dk.kb.hadoop.nark.CDXJob \
    TEST1_list.txt \
    outputFolder

inputfile is a test file consisting of warc-file addresses. Remember to use the prefix file:// as the addresses are resolved on HDFS per default. This goes both for the inputfile and the lines in inputfile.

Note that inputfile and outputFolder must be available on all nodes in the cluster, so it is probably wise to use HDFS for these.

Get the result like this

hdfs dfs -ls outputFolder/
hdfs dfs -cat 'outputFolder/part-r-*'

Run this on the server [email protected]. The netarchive files are available on /mnt/netarkiv/

Generate input file

ssh [email protected]
find  /mnt/netarkiv/ -type f | sed 's|\(.*\)|file://\1|' > TEST1_list.txt
hdfs dfs -copyFromLocal TEST1_list.txt .