Collecting stashcp information

The stashcp command has been instrumented to send the most important information in a json format to a flume collector running at University of Chicago.

The flume instance is configured to use http source listening on hadoop-dev.mwt2.org:80. It "sinks" the received data into:

• log files(/var/log/StashCollector)
• HDFS (/user/ivukotic/StashCollector/)
• ElasticSearch(index name: stashcache).

Flume instance can be monitored at: http://hadoop-dev.mwt2.org:34548/metrics