You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mesos.apache.org by be...@apache.org on 2011/06/05 11:03:25 UTC

svn commit: r1132222 - in /incubator/mesos/trunk/ec2: deploy.centos64/root/ephemeral-hdfs/conf/ deploy.centos64/root/mesos-ec2/ deploy.centos64/root/mesos-ec2/hadoop-framework-conf/ deploy.lucid64/root/mesos-ec2/

Author: benh
Date: Sun Jun  5 09:03:25 2011
New Revision: 1132222

URL: http://svn.apache.org/viewvc?rev=1132222&view=rev
Log:
A few EC2 fixes:

- Decreased some of the default settings for Hadoop on EC2 to work better
  on smaller nodes.

- Increased timeout in start-mesos to make sure master is up before
  slaves try to connect to it.

- Fixed a potential infinite retry cycle in Doug's code that would
  happen if a node was unavailable (e.g. was terminated in ElasticFox)

Modified:
    incubator/mesos/trunk/ec2/deploy.centos64/root/ephemeral-hdfs/conf/hadoop-env.sh
    incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/hadoop-env.sh
    incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/mapred-site.xml
    incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/setup
    incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/start-mesos
    incubator/mesos/trunk/ec2/deploy.lucid64/root/mesos-ec2/start-mesos

Modified: incubator/mesos/trunk/ec2/deploy.centos64/root/ephemeral-hdfs/conf/hadoop-env.sh
URL: http://svn.apache.org/viewvc/incubator/mesos/trunk/ec2/deploy.centos64/root/ephemeral-hdfs/conf/hadoop-env.sh?rev=1132222&r1=1132221&r2=1132222&view=diff
==============================================================================
--- incubator/mesos/trunk/ec2/deploy.centos64/root/ephemeral-hdfs/conf/hadoop-env.sh (original)
+++ incubator/mesos/trunk/ec2/deploy.centos64/root/ephemeral-hdfs/conf/hadoop-env.sh Sun Jun  5 09:03:25 2011
@@ -12,7 +12,7 @@ export JAVA_HOME=/usr/java/default
 # export HADOOP_CLASSPATH=
 
 # The maximum amount of heap to use, in MB. Default is 1000.
-export HADOOP_HEAPSIZE=1500
+export HADOOP_HEAPSIZE=1000
 
 # Extra Java runtime options.  Empty by default.
 # export HADOOP_OPTS=-server

Modified: incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/hadoop-env.sh
URL: http://svn.apache.org/viewvc/incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/hadoop-env.sh?rev=1132222&r1=1132221&r2=1132222&view=diff
==============================================================================
--- incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/hadoop-env.sh (original)
+++ incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/hadoop-env.sh Sun Jun  5 09:03:25 2011
@@ -12,7 +12,7 @@ export JAVA_HOME=/usr/java/default
 # export HADOOP_CLASSPATH=
 
 # The maximum amount of heap to use, in MB. Default is 1000.
-export HADOOP_HEAPSIZE=1500
+export HADOOP_HEAPSIZE=1000
 
 # Extra Java runtime options.  Empty by default.
 # export HADOOP_OPTS=-server

Modified: incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/mapred-site.xml
URL: http://svn.apache.org/viewvc/incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/mapred-site.xml?rev=1132222&r1=1132221&r2=1132222&view=diff
==============================================================================
--- incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/mapred-site.xml (original)
+++ incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/hadoop-framework-conf/mapred-site.xml Sun Jun  5 09:03:25 2011
@@ -37,12 +37,17 @@
 
   <property>
     <name>tasktracker.http.threads</name>
-    <value>50</value>
+    <value>40</value>
   </property>
 
   <property>
     <name>mapred.child.java.opts</name>
-    <value>-Xmx1512m</value>
+    <value>-Xmx500m</value>
+  </property>
+
+  <property>
+    <name>mapred.mesos.task.mem</name>
+    <value>500</value>
   </property>
 
   <property>
@@ -57,7 +62,7 @@
 
   <property>
     <name>io.sort.mb</name>
-    <value>100</value>
+    <value>150</value>
   </property>
 
   <property>
@@ -75,12 +80,4 @@
     <value>8</value>
   </property>
 
-  <property>
-    <name>io.sort.mb</name>
-    <value>200</value>
-    <description>The total amount of buffer memory to use while sorting
-    files, in megabytes.  By default, gives each merge stream 1MB, which
-    should minimize seeks.</description>
-  </property>
-
 </configuration>

Modified: incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/setup
URL: http://svn.apache.org/viewvc/incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/setup?rev=1132222&r1=1132221&r2=1132222&view=diff
==============================================================================
--- incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/setup (original)
+++ incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/setup Sun Jun  5 09:03:25 2011
@@ -82,9 +82,12 @@ if [[ $NUM_ZOOS != 0 ]] ; then
   done
 fi
 
-TODO="$SLAVES"
+# Try to SSH to each slave to approve their key. Since some slaves may be
+# slow in starting, we retry failed slaves up to 3 times.
+TODO="$SLAVES"  # List of slaves to try (initially all of them)
+TRIES="0"       # Number of times we've tried so far
 echo "SSH'ing to slaves to approve keys..."
-while [ "e$TODO" != "e" ] ; do
+while [ "e$TODO" != "e" ] && [ $TRIES -lt 4 ] ; do
   NEW_TODO=
   for slave in $TODO; do
     echo $slave
@@ -93,8 +96,9 @@ while [ "e$TODO" != "e" ] ; do
         NEW_TODO="$NEW_TODO $slave"
     fi
   done
-  if [ "e$NEW_TODO" != "e" ] ; then
-      sleep 10
+  TRIES=$[$TRIES + 1]
+  if [ "e$NEW_TODO" != "e" ] && [ $TRIES -lt 4 ] ; then
+      sleep 15
       TODO="$NEW_TODO"
       echo "Re-attempting SSH to slaves to approve keys..."
   else

Modified: incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/start-mesos
URL: http://svn.apache.org/viewvc/incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/start-mesos?rev=1132222&r1=1132221&r2=1132222&view=diff
==============================================================================
--- incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/start-mesos (original)
+++ incubator/mesos/trunk/ec2/deploy.centos64/root/mesos-ec2/start-mesos Sun Jun  5 09:03:25 2011
@@ -32,7 +32,7 @@ else
   ssh $SSH_OPTS $ACTIVE_MASTER "/root/mesos-ec2/mesos-daemon mesos-master -p 5050 $@ </dev/null >/dev/null"
 fi
 
-sleep 2
+sleep 5
 
 for slave in $SLAVES; do
   echo "Starting slave on $slave"

Modified: incubator/mesos/trunk/ec2/deploy.lucid64/root/mesos-ec2/start-mesos
URL: http://svn.apache.org/viewvc/incubator/mesos/trunk/ec2/deploy.lucid64/root/mesos-ec2/start-mesos?rev=1132222&r1=1132221&r2=1132222&view=diff
==============================================================================
--- incubator/mesos/trunk/ec2/deploy.lucid64/root/mesos-ec2/start-mesos (original)
+++ incubator/mesos/trunk/ec2/deploy.lucid64/root/mesos-ec2/start-mesos Sun Jun  5 09:03:25 2011
@@ -32,7 +32,7 @@ else
   ssh $SSH_OPTS $ACTIVE_MASTER "/root/mesos-ec2/mesos-daemon mesos-master -p 5050 $@ </dev/null >/dev/null"
 fi
 
-sleep 2
+sleep 5
 
 for slave in $SLAVES; do
   echo "Starting slave on $slave"