From e756226ab91e110a22db3fee33e2f3df46da300e Mon Sep 17 00:00:00 2001
From: Shahrukh Khan <shahrukhkhan489@gmail.com>
Date: Fri, 10 Nov 2017 19:46:46 +0530
Subject: [PATCH] TUNING Updating default MR fetcher for performance (#300)

---
 app-conf/FetcherConf.xml | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/app-conf/FetcherConf.xml b/app-conf/FetcherConf.xml
index 4e9f14df8..f81aad5e6 100644
--- a/app-conf/FetcherConf.xml
+++ b/app-conf/FetcherConf.xml
@@ -29,6 +29,7 @@
   </fetcher>
 -->
 <fetchers>
+  <!--
   <fetcher>
     <applicationtype>mapreduce</applicationtype>
     <classname>com.linkedin.drelephant.mapreduce.fetchers.MapReduceFetcherHadoop2</classname>
@@ -36,20 +37,21 @@
       <sampling_enabled>false</sampling_enabled>
     </params>
   </fetcher>
+  -->
   <!--
-     This is an experimental replacement for the MapReduceFetcherHadoop2 that attempts to burn
+     This is a replacement for the MapReduceFetcherHadoop2 that attempts to burn
      through queues of jobs faster by pulling data directly from HDFS rather than going through
      the job history server.
 
      Increasing the param history_log_size_limit_in_mb allows this fetcher to accept larger log
      files, but also increase the risk of OutOfMemory error. The default heap size of Dr. Elephant
-     is 1024MB. To increase this, e.g. to 2048MB, run this before start.sh:
-       export OPTS="-mem 2048"
+     is 1024MB. To increase this, e.g. to 2048MB, update the below mem conf in app-conf/elephant.conf:
+       jvm_args="-mem 2048"
 
      To work properly, this fetcher should use the same timezone with the job history server.
      If not set, the local timezone will be used.
    -->
-  <!--
+  
   <fetcher>
     <applicationtype>mapreduce</applicationtype>
     <classname>com.linkedin.drelephant.mapreduce.fetchers.MapReduceFSFetcherHadoop2</classname>
@@ -59,7 +61,7 @@
       <history_server_time_zone>PST</history_server_time_zone>
     </params>
   </fetcher>
-  -->
+  
 
   <!--
     FSFetcher for Spark. Loads the eventlog from HDFS and replays to get the metrics and application properties