Disabling Transparent Hugepage Compaction

Most Linux platforms supported by CDH 5 include a feature called transparent hugepage compaction which interacts poorly with Hadoop workloads and can seriously degrade performance.

Symptom: top and other system monitoring tools show a large percentage of the CPU usage classified as "system CPU". If system CPU usage is 30% or more of the total CPU usage, your system may be experiencing this issue.

What to do:

  Note: In the following instructions, defrag_file_pathname depends on your operating system:

• Red Hat/CentOS: /sys/kernel/mm/redhat_transparent_hugepage/defrag
• Ubuntu/Debian, OEL, SLES: /sys/kernel/mm/transparent_hugepage/defrag

1. To see whether transparent hugepage compaction is enabled, run the following command and check the output:

   $ cat defrag_file_pathname

   • [always] never means that transparent hugepage compaction is enabled.
   • always [never] means that transparent hugepage compaction is disabled.
2. To disable transparent hugepage compaction, add the following command to /etc/rc.local:

    echo never > defrag_file_pathname

You can also disable transparent hugepage compaction interactively (but remember this will not survive a reboot).

# echo 'never' > defrag_file_pathname 

$ sudo sh -c "echo 'never' > defrag_file_pathname" 

Setting the vm.swappiness Linux Kernel Parameter

The Linux kernel parameter, vm.swappiness, is a value from 0-100 that controls the swapping of application data (as anonymous pages) from physical memory to virtual memory on disk. The higher the value, the more aggressively inactive processes are swapped out from physical memory. The lower the value, the less they are swapped, forcing filesystem buffers to be emptied.

On most systems, vm.swappiness is set to 60 by default. This is not suitable for Hadoop clusters because processes are sometimes swapped even when enough memory is available. This can cause lengthy garbage collection pauses for important system daemons, affecting stability and performance.

Cloudera recommends that you set vm.swappiness to a value between 1 and 10, preferably 1, for minimum swapping.

cat /proc/sys/vm/swappiness

sudo sysctl -w vm.swappiness=1

  Note: Cloudera previously recommended setting vm.swappiness to 0. However, a change in Linux kernel 3.5-rc1 (fe35004f), can lead to frequent out of memory (OOM) errors. For details, see Evan Klitzke's blog post. This commit was backported to RHEL / CentOS 6.4 and Ubuntu 12.04 LTS (Long Term Support).

Improving Performance in Shuffle Handler and IFile Reader

The MapReduce shuffle handler and IFile reader use native Linux calls, (posix_fadvise(2) and sync_data_range), on Linux systems with Hadoop native libraries installed.

Shuffle Handler

You can improve MapReduce shuffle handler performance by enabling shuffle readahead. This causes the TaskTracker or Node Manager to pre-fetch map output before sending it over the socket to the reducer.

• To enable this feature for YARN, set mapreduce.shuffle.manage.os.cache, to true (default). To further tune performance, adjust the value of mapreduce.shuffle.readahead.bytes. The default value is 4 MB.
• To enable this feature for MapReduce, set the mapred.tasktracker.shuffle.fadvise to true (default). To further tune performance, adjust the value of mapred.tasktracker.shuffle.readahead.bytes. The default value is 4 MB.

IFile Reader

Enabling IFile readahead increases the performance of merge operations. To enable this feature for either MRv1 or YARN, set mapreduce.ifile.readahead to true (default). To further tune the performance, adjust the value of mapreduce.ifile.readahead.bytes. The default value is 4MB.

Best Practices for MapReduce Configuration

The configuration settings described below can reduce inherent latencies in MapReduce execution. You set these values in mapred-site.xml.

Send a heartbeat as soon as a task finishes

Set mapreduce.tasktracker.outofband.heartbeat to true for TaskTracker to send an out-of-band heartbeat on task completion to reduce latency. The default value is false:

<property>
    <name>mapreduce.tasktracker.outofband.heartbeat</name>
    <value>true</value>
</property>

Reduce the interval for JobClient status reports on single node systems

The jobclient.progress.monitor.poll.interval property defines the interval (in milliseconds) at which JobClient reports status to the console and checks for job completion. The default value is 1000 milliseconds; you may want to set this to a lower value to make tests run faster on a single-node cluster. Adjusting this value on a large production cluster may lead to unwanted client-server traffic.

<property>
    <name>jobclient.progress.monitor.poll.interval</name>
    <value>10</value>
</property>

Tune the JobTracker heartbeat interval

Tuning the minimum interval for the TaskTracker-to-JobTracker heartbeat to a smaller value may improve MapReduce performance on small clusters.

<property>
    <name>mapreduce.jobtracker.heartbeat.interval.min</name>
    <value>10</value>
</property>

Start MapReduce JVMs immediately

The mapred.reduce.slowstart.completed.maps property specifies the proportion of Map tasks in a job that must be completed before any Reduce tasks are scheduled. For small jobs that require fast turnaround, setting this value to 0 can improve performance; larger values (as high as 50%) may be appropriate for larger jobs.

<property>
    <name>mapred.reduce.slowstart.completed.maps</name>
    <value>0</value>
</property>

Tips and Best Practices for Jobs

This section describes changes you can make at the job level.

Use the Distributed Cache to Transfer the Job JAR

Use the distributed cache to transfer the job JAR rather than using the JobConf(Class) constructor and the JobConf.setJar() and JobConf.setJarByClass() methods.

To add JARs to the classpath, use -libjars jar1,jar2. This copies the local JAR files to HDFS and uses the distributed cache mechanism to ensure they are available on the task nodes and added to the task classpath.

The advantage of this, over JobConf.setJar, is that if the JAR is on a task node, it does not need to be copied again if a second task from the same job runs on that node, though it will still need to be copied from the launch machine to HDFS.

  Note: -libjars works only if your MapReduce driver uses ToolRunner. If it does not, you would need to use the DistributedCache APIs (Cloudera does not recommend this).

For more information, see item 1 in the blog post How to Include Third-Party Libraries in Your MapReduce Job.

Changing the Logging Level on a Job (MRv1)

You can change the logging level for an individual job. You do this by setting the following properties in the job configuration (JobConf):

• mapreduce.map.log.level
• mapreduce.reduce.log.level

Valid values are NONE, INFO, WARN, DEBUG, TRACE, and ALL.

Example:

JobConf conf = new JobConf();
...

conf.set("mapreduce.map.log.level", "DEBUG");
conf.set("mapreduce.reduce.log.level", "TRACE");
...