From 10c7f0eb1526b8d66b79138d387243746d85e0f5 Mon Sep 17 00:00:00 2001
From: Konstantin Shvachko <shv@apache.org>
Date: Mon, 13 Feb 2012 21:16:22 +0000
Subject: [PATCH] MAPREDUCE-3837. Job tracker is not able to recover jobs after
 crash. Contributed by Mayank Bansal.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-0.23@1243698 13f79535-47bb-0310-9956-ffa450edef68
---
 hadoop-mapreduce-project/CHANGES.txt               |  3 +++
 .../java/org/apache/hadoop/mapred/JobTracker.java  | 14 +++++++++-----
 2 files changed, 12 insertions(+), 5 deletions(-)

diff --git a/hadoop-mapreduce-project/CHANGES.txt b/hadoop-mapreduce-project/CHANGES.txt
index c045b0f1100..d084f61278e 100644
--- a/hadoop-mapreduce-project/CHANGES.txt
+++ b/hadoop-mapreduce-project/CHANGES.txt
@@ -2649,6 +2649,9 @@ Release 0.22.1 - Unreleased
 
   BUG FIXES
 
+    MAPREDUCE-3837. Job tracker is not able to recover jobs after crash.
+    (Mayank Bansal via shv) 
+
 Release 0.22.0 - 2011-11-29
 
   INCOMPATIBLE CHANGES
diff --git a/hadoop-mapreduce-project/src/java/org/apache/hadoop/mapred/JobTracker.java b/hadoop-mapreduce-project/src/java/org/apache/hadoop/mapred/JobTracker.java
index 239038f707e..8a07128cc78 100644
--- a/hadoop-mapreduce-project/src/java/org/apache/hadoop/mapred/JobTracker.java
+++ b/hadoop-mapreduce-project/src/java/org/apache/hadoop/mapred/JobTracker.java
@@ -1192,13 +1192,17 @@ public class JobTracker implements MRConstants, InterTrackerProtocol,
         try {
           Path jobInfoFile = getSystemFileForJob(jobId);
           FSDataInputStream in = fs.open(jobInfoFile);
-          JobInfo token = new JobInfo();
+          final JobInfo token = new JobInfo();
           token.readFields(in);
           in.close();
-          UserGroupInformation ugi = 
-            UserGroupInformation.createRemoteUser(token.getUser().toString());
-          submitJob(token.getJobID(), restartCount, 
-              ugi, token.getJobSubmitDir().toString(), true, null);
+          final UserGroupInformation ugi = 
+          UserGroupInformation.createRemoteUser(token.getUser().toString());
+          ugi.doAs(new PrivilegedExceptionAction<JobStatus>() {
+            public JobStatus run() throws IOException ,InterruptedException{
+               return submitJob(token.getJobID(), restartCount, 
+                        ugi, token.getJobSubmitDir().toString(), true, null);
+          }});
+          
           recovered++;
         } catch (Exception e) {
           LOG.warn("Could not recover job " + jobId, e);