Skip to content

SPARK-1818 Freshen Mesos documentation #756

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 9 commits into from
Closed

SPARK-1818 Freshen Mesos documentation #756

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
74f2040
SPARK-1818 Freshen Mesos documentation
ash211 May 13, 2014
8bf2c53
Update site.MESOS_VERSIOn to match /pom.xml
ash211 May 13, 2014
41b68ed
Period at end of sentence; formatting on :5050
ash211 May 13, 2014
3dcc2c1
Use --tgz parameter of make-distribution
ash211 May 13, 2014
7ebf6ef
Polish on the section on Mesos Master URLs
ash211 May 13, 2014
de3353b
Wrap to 100char
ash211 May 13, 2014
956362d
Don't need to pass spark.executor.uri into the spark shell
ash211 May 13, 2014
e7dea8e
Add troubleshooting and debugging section
ash211 May 14, 2014
7ef3b33
Brief explanation of the interactions between Spark and Mesos
ash211 May 14, 2014
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion docs/_config.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,6 @@ SPARK_VERSION: 1.0.0-SNAPSHOT
SPARK_VERSION_SHORT: 1.0.0
SCALA_BINARY_VERSION: "2.10"
SCALA_VERSION: "2.10.4"
MESOS_VERSION: 0.13.0
MESOS_VERSION: 0.18.1
SPARK_ISSUE_TRACKER_URL: https://issues.apache.org/jira/browse/SPARK
SPARK_GITHUB_URL: https://github.com/apache/spark
200 changes: 173 additions & 27 deletions docs/running-on-mesos.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,19 +3,123 @@ layout: global
title: Running Spark on Mesos
---

Spark can run on clusters managed by [Apache Mesos](http://mesos.apache.org/). Follow the steps below to install Mesos and Spark:

1. Download and build Spark using the instructions [here](index.html). **Note:** Don't forget to consider what version of HDFS you might want to use!
2. Download, build, install, and start Mesos {{site.MESOS_VERSION}} on your cluster. You can download the Mesos distribution from a [mirror](http://www.apache.org/dyn/closer.cgi/mesos/{{site.MESOS_VERSION}}/). See the Mesos [Getting Started](http://mesos.apache.org/gettingstarted) page for more information. **Note:** If you want to run Mesos without installing it into the default paths on your system (e.g., if you don't have administrative privileges to install it), you should also pass the `--prefix` option to `configure` to tell it where to install. For example, pass `--prefix=/home/user/mesos`. By default the prefix is `/usr/local`.
3. Create a Spark "distribution" using `make-distribution.sh`.
4. Rename the `dist` directory created from `make-distribution.sh` to `spark-{{site.SPARK_VERSION}}`.
5. Create a `tar` archive: `tar czf spark-{{site.SPARK_VERSION}}.tar.gz spark-{{site.SPARK_VERSION}}`
6. Upload this archive to HDFS or another place accessible from Mesos via `http://`, e.g., [Amazon Simple Storage Service](http://aws.amazon.com/s3): `hadoop fs -put spark-{{site.SPARK_VERSION}}.tar.gz /path/to/spark-{{site.SPARK_VERSION}}.tar.gz`
7. Create a file called `spark-env.sh` in Spark's `conf` directory, by copying `conf/spark-env.sh.template`, and add the following lines to it:
* `export MESOS_NATIVE_LIBRARY=<path to libmesos.so>`. This path is usually `<prefix>/lib/libmesos.so` (where the prefix is `/usr/local` by default, see above). Also, on Mac OS X, the library is called `libmesos.dylib` instead of `libmesos.so`.
* `export SPARK_EXECUTOR_URI=<path to spark-{{site.SPARK_VERSION}}.tar.gz uploaded above>`.
* `export MASTER=mesos://HOST:PORT` where HOST:PORT is the host and port (default: 5050) of your Mesos master (or `zk://...` if using Mesos with ZooKeeper).
8. To run a Spark application against the cluster, when you create your `SparkContext`, pass the string `mesos://HOST:PORT` as the master URL. In addition, you'll need to set the `spark.executor.uri` property. For example:
# Why Mesos
Copy link

@maasg maasg May 13, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I miss a chapter on "How Spark works on Mesos" that helps understanding the need of the assembly in a reachable location by the slaves and how the spark driver program becomes the 'scheduler' in Mesos terms and the tasks get delivered to the executors.


Spark can run on hardware clusters managed by [Apache Mesos](http://mesos.apache.org/).

The advantages of deploying Spark with Mesos include:
- dynamic partitioning between Spark and other
[frameworks](https://mesos.apache.org/documentation/latest/mesos-frameworks/)
- scalable partitioning between multiple instances of Spark

# How it works

In a standalone cluster deployment, the cluster manager in the below diagram is a Spark master
instance. When using Mesos, the Mesos master replaces the Spark master as the cluster manager.

<p style="text-align: center;">
<img src="img/cluster-overview.png" title="Spark cluster components" alt="Spark cluster components" />
</p>

Now when a driver creates a job and starts issuing tasks for scheduling, Mesos determines what
machines handle what tasks. Because it takes into account other frameworks when scheduling these
many short-lived tasks, multiple frameworks can coexist on the same cluster without resorting to a
static partitioning of resources.

To get started, follow the steps below to install Mesos and deploy Spark jobs via Mesos.


# Installing Mesos

Spark {{site.SPARK_VERSION}} is designed for use with Mesos {{site.MESOS_VERSION}} and does not
Copy link
Contributor

@pwendell pwendell May 13, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Incidentally I think MESOS_VERSION needs to be updated to 0.18 - mind doing that?

Copy link
Contributor Author

@ash211 ash211 May 13, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Updated to 0.18.1 per the pom.xml

require any special patches of Mesos.

If you already have a Mesos cluster running, you can skip this Mesos installation step.

Otherwise, installing Mesos for Spark is no different than installing Mesos for use by other
frameworks. You can install Mesos using either prebuilt packages or by compiling from source.

## Prebuilt packages

The Apache Mesos project only publishes source package releases, no binary releases. But other
third party projects publish binary releases that may be helpful in setting Mesos up.

One of those is Mesosphere. To install Mesos using the binary releases provided by Mesosphere:
Copy link
Contributor

@mateiz mateiz May 16, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ash211 @pwendell, I missed this when the PR was being reviewed, but we need to list the source instructions first. We are first and foremost an Apache project, and need to tell people how to use Apache software to get their work done. Third parties can go ahead and publish binary distributions if they want, but we can't be posting that as the preferred way to install another Apache project. We do the same thing for Hadoop distros -- we say how to build against an Apache release of Hadoop, and then we have a page on building against third-party distros.

Copy link
Contributor

@mateiz mateiz May 16, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also I would call these "third-party packages" instead of "prebuilt packages". Again, since this is going on an Apache web page, we can't be misrepresenting the Mesos project. Bits that are not built by Apache are unfortunately third-party.

Copy link
Contributor Author

@ash211 ash211 May 16, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Got it. I'm starting a cleanup PR that should be in within an hour.

On Fri, May 16, 2014 at 3:17 PM, Matei Zaharia [email protected]:

In docs/running-on-mesos.md:

+# Installing Mesos
+
+Spark {{site.SPARK_VERSION}} is designed for use with Mesos {{site.MESOS_VERSION}} and does not
+require any special patches of Mesos.
+
+If you already have a Mesos cluster running, you can skip this Mesos installation step.
+
+Otherwise, installing Mesos for Spark is no different than installing Mesos for use by other
+frameworks. You can install Mesos using either prebuilt packages or by compiling from source.
+
+## Prebuilt packages
+
+The Apache Mesos project only publishes source package releases, no binary releases. But other
+third party projects publish binary releases that may be helpful in setting Mesos up.
+
+One of those is Mesosphere. To install Mesos using the binary releases provided by Mesosphere:

Also I would call these "third-party packages" instead of "prebuilt
packages". Again, since this is going on an Apache web page, we can't be
misrepresenting the Mesos project. Bits that are not built by Apache are
unfortunately third-party.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/756/files#r12761962
.

Copy link
Contributor Author

@ash211 ash211 May 16, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

#805

On Fri, May 16, 2014 at 3:41 PM, Andrew Ash [email protected] wrote:

Got it. I'm starting a cleanup PR that should be in within an hour.

On Fri, May 16, 2014 at 3:17 PM, Matei Zaharia [email protected]:

In docs/running-on-mesos.md:

+# Installing Mesos
+
+Spark {{site.SPARK_VERSION}} is designed for use with Mesos {{site.MESOS_VERSION}} and does not
+require any special patches of Mesos.
+
+If you already have a Mesos cluster running, you can skip this Mesos installation step.
+
+Otherwise, installing Mesos for Spark is no different than installing Mesos for use by other
+frameworks. You can install Mesos using either prebuilt packages or by compiling from source.
+
+## Prebuilt packages
+
+The Apache Mesos project only publishes source package releases, no binary releases. But other
+third party projects publish binary releases that may be helpful in setting Mesos up.
+
+One of those is Mesosphere. To install Mesos using the binary releases provided by Mesosphere:

Also I would call these "third-party packages" instead of "prebuilt
packages". Again, since this is going on an Apache web page, we can't be
misrepresenting the Mesos project. Bits that are not built by Apache are
unfortunately third-party.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/756/files#r12761962
.


1. Download Mesos installation package from [downloads page](http://mesosphere.io/downloads/)
2. Follow their instructions for installation and configuration

The Mesosphere installation documents suggest setting up ZooKeeper to handle Mesos master failover,
but Mesos can be run without ZooKeeper using a single master as well.

## From source

To install Mesos directly from the upstream project rather than a third party, install from source.

1. Download the Mesos distribution from a
[mirror](http://www.apache.org/dyn/closer.cgi/mesos/{{site.MESOS_VERSION}}/)
2. Follow the Mesos [Getting Started](http://mesos.apache.org/gettingstarted) page for compiling and
installing Mesos

**Note:** If you want to run Mesos without installing it into the default paths on your system
(e.g., if you lack administrative privileges to install it), you should also pass the
`--prefix` option to `configure` to tell it where to install. For example, pass
`--prefix=/home/user/mesos`. By default the prefix is `/usr/local`.

## Verification

To verify that the Mesos cluster is ready for Spark, navigate to the Mesos master webui at port
`:5050` Confirm that all expected machines are present in the slaves tab.


# Connecting Spark to Mesos

To use Mesos from Spark, you need a Spark distribution available in a place accessible by Mesos, and
a Spark driver program configured to connect to Mesos.

## Uploading Spark Distribution

When Mesos runs a task on a Mesos slave for the first time, that slave must have a distribution of
Spark available for running the Spark Mesos executor backend. A distribution of Spark is just a
compiled binary version of Spark.

The Spark distribution can be hosted at any Hadoop URI, including HTTP via `http://`, [Amazon Simple
Storage Service](http://aws.amazon.com/s3) via `s3://`, or HDFS via `hdfs:///`.

To use a precompiled distribution:

1. Download a Spark distribution from the Spark [download page](https://spark.apache.org/downloads.html)
2. Upload to hdfs/http/s3

To host on HDFS, use the Hadoop fs put command: `hadoop fs -put spark-{{site.SPARK_VERSION}}.tar.gz
/path/to/spark-{{site.SPARK_VERSION}}.tar.gz`


Or if you are using a custom-compiled version of Spark, you will need to create a distribution using
the `make-distribution.sh` script included in a Spark source tarball/checkout.

1. Download and build Spark using the instructions [here](index.html)
2. Create a Spark distribution using `make-distribution.sh --tgz`.
3. Upload archive to http/s3/hdfs


## Using a Mesos Master URL

The Master URLs for Mesos are in the form `mesos://host:5050` for a single-master Mesos
cluster, or `zk://host:2181` for a multi-master Mesos cluster using ZooKeeper.

The driver also needs some configuration in `spark-env.sh` to interact properly with Mesos:

1. In `spark.env.sh` set some environment variables:
* `export MESOS_NATIVE_LIBRARY=<path to libmesos.so>`. This path is typically
`<prefix>/lib/libmesos.so` where the prefix is `/usr/local` by default. See Mesos installation
instructions above. On Mac OS X, the library is called `libmesos.dylib` instead of
`libmesos.so`.
* `export SPARK_EXECUTOR_URI=<path to spark-{{site.SPARK_VERSION}}.tar.gz uploaded above>`.
2. Also set `spark.executor.uri` to <path to spark-{{site.SPARK_VERSION}}.tar.gz>

Now when starting a Spark application against the cluster, pass a `mesos://`
or `zk://` URL as the master when creating a `SparkContext`. For example:

{% highlight scala %}
val conf = new SparkConf()
Expand All @@ -25,31 +129,73 @@ val conf = new SparkConf()
val sc = new SparkContext(conf)
{% endhighlight %}

When running a shell the `spark.executor.uri` parameter is inherited from `SPARK_EXECUTOR_URI`, so
it does not need to be redundantly passed in as a system property.

{% highlight shell %}
Copy link
Contributor

@pwendell pwendell May 14, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this needs to be highlight bash. At least on my version of jekyll I get an error otherwise when compiling docs.

./bin/spark-shell --master mesos://host:5050
{% endhighlight %}


# Mesos Run Modes

Spark can run over Mesos in two modes: "fine-grained" and "coarse-grained". In fine-grained mode, which is the default,
each Spark task runs as a separate Mesos task. This allows multiple instances of Spark (and other frameworks) to share
machines at a very fine granularity, where each application gets more or fewer machines as it ramps up, but it comes with an
additional overhead in launching each task, which may be inappropriate for low-latency applications (e.g. interactive queries or serving web requests). The coarse-grained mode will instead
launch only *one* long-running Spark task on each Mesos machine, and dynamically schedule its own "mini-tasks" within
it. The benefit is much lower startup overhead, but at the cost of reserving the Mesos resources for the complete duration
of the application.
Spark can run over Mesos in two modes: "fine-grained" (default) and "coarse-grained".

In "fine-grained" mode (default), each Spark task runs as a separate Mesos task. This allows
multiple instances of Spark (and other frameworks) to share machines at a very fine granularity,
where each application gets more or fewer machines as it ramps up and down, but it comes with an
additional overhead in launching each task. This mode may be inappropriate for low-latency
requirements like interactive queries or serving web requests.

The "coarse-grained" mode will instead launch only *one* long-running Spark task on each Mesos
machine, and dynamically schedule its own "mini-tasks" within it. The benefit is much lower startup
overhead, but at the cost of reserving the Mesos resources for the complete duration of the
application.

To run in coarse-grained mode, set the `spark.mesos.coarse` property in your [SparkConf](configuration.html#spark-properties):
To run in coarse-grained mode, set the `spark.mesos.coarse` property in your
[SparkConf](configuration.html#spark-properties):

{% highlight scala %}
conf.set("spark.mesos.coarse", "true")
{% endhighlight %}

In addition, for coarse-grained mode, you can control the maximum number of resources Spark will acquire. By default,
it will acquire *all* cores in the cluster (that get offered by Mesos), which only makes sense if you run just one
application at a time. You can cap the maximum number of cores using `conf.set("spark.cores.max", "10")` (for example).
In addition, for coarse-grained mode, you can control the maximum number of resources Spark will
acquire. By default, it will acquire *all* cores in the cluster (that get offered by Mesos), which
only makes sense if you run just one application at a time. You can cap the maximum number of cores
using `conf.set("spark.cores.max", "10")` (for example).


Copy link

@maasg maasg May 13, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Troubleshooting and debugging

  • Accessing the tasks on the mesos console
  • Location of the logs on the Mesos executor
  • Usual pitfalls
    -- Spark assembly is not reachable/accessible
    -- Slaves cannot reach resources (HDFS/...)
    -- Tasks cannot reach driver (when e.g. your try to submit from a local machine to a remote cluster)

...

# Running Alongside Hadoop

You can run Spark and Mesos alongside your existing Hadoop cluster by just launching them as a separate service on the machines. To access Hadoop data from Spark, just use a hdfs:// URL (typically `hdfs://<namenode>:9000/path`, but you can find the right URL on your Hadoop Namenode's web UI).
You can run Spark and Mesos alongside your existing Hadoop cluster by just launching them as a
separate service on the machines. To access Hadoop data from Spark, a full hdfs:// URL is required
(typically `hdfs://<namenode>:9000/path`, but you can find the right URL on your Hadoop Namenode web
UI).

In addition, it is possible to also run Hadoop MapReduce on Mesos for better resource isolation and
sharing between the two. In this case, Mesos will act as a unified scheduler that assigns cores to
either Hadoop or Spark, as opposed to having them share resources via the Linux scheduler on each
node. Please refer to [Hadoop on Mesos](https://github.com/mesos/hadoop).

In either case, HDFS runs separately from Hadoop MapReduce, without being scheduled through Mesos.


# Troubleshooting and Debugging

A few places to look during debugging:

- Mesos master on port `:5050`
- Slaves should appear in the slaves tab
- Spark applications should appear in the frameworks tab
- Tasks should appear in the details of a framework
- Check the stdout and stderr of the sandbox of failed tasks
- Mesos logs
- Master and slave logs are both in `/var/log/mesos` by default

In addition, it is possible to also run Hadoop MapReduce on Mesos, to get better resource isolation and sharing between the two. In this case, Mesos will act as a unified scheduler that assigns cores to either Hadoop or Spark, as opposed to having them share resources via the Linux scheduler on each node. Please refer to [Hadoop on Mesos](https://github.com/mesos/hadoop).
And common pitfalls:

In either case, HDFS runs separately from Hadoop MapReduce, without going through Mesos.
- Spark assembly not reachable/accessible
- Slaves need to be able to download the distribution
- Firewall blocking communications
- Check for messages about failed connections
- Temporarily disable firewalls for debugging and then poke appropriate holes