Revision f2408bdd7a0950385ee1364e006d55bfa6e5a200 authored by Shixiong Zhu on 27 May 2017, 05:25:38 UTC, committed by Shixiong Zhu on 27 May 2017, 05:25:44 UTC
## What changes were proposed in this pull request? Add a `worker` configuration to set how long to wait before forcibly killing driver. ## How was this patch tested? Jenkins Author: Shixiong Zhu <shixiong@databricks.com> Closes #18126 from zsxwing/SPARK-20843. (cherry picked from commit 6c1dbd6fc8d49acf7c1c902d2ebf89ed5e788a4e) Signed-off-by: Shixiong Zhu <shixiong@databricks.com>
1 parent 39f7665
storage-openstack-swift.md
---
layout: global
title: Accessing OpenStack Swift from Spark
---
Spark's support for Hadoop InputFormat allows it to process data in OpenStack Swift using the
same URI formats as in Hadoop. You can specify a path in Swift as input through a
URI of the form <code>swift://container.PROVIDER/path</code>. You will also need to set your
Swift security credentials, through <code>core-site.xml</code> or via
<code>SparkContext.hadoopConfiguration</code>.
Current Swift driver requires Swift to use Keystone authentication method.
# Configuring Swift for Better Data Locality
Although not mandatory, it is recommended to configure the proxy server of Swift with
<code>list_endpoints</code> to have better data locality. More information is
[available here](https://github.com/openstack/swift/blob/master/swift/common/middleware/list_endpoints.py).
# Dependencies
The Spark application should include <code>hadoop-openstack</code> dependency.
For example, for Maven support, add the following to the <code>pom.xml</code> file:
{% highlight xml %}
<dependencyManagement>
...
<dependency>
<groupId>org.apache.hadoop</groupId>
<artifactId>hadoop-openstack</artifactId>
<version>2.3.0</version>
</dependency>
...
</dependencyManagement>
{% endhighlight %}
# Configuration Parameters
Create <code>core-site.xml</code> and place it inside Spark's <code>conf</code> directory.
There are two main categories of parameters that should to be configured: declaration of the
Swift driver and the parameters that are required by Keystone.
Configuration of Hadoop to use Swift File system achieved via
<table class="table">
<tr><th>Property Name</th><th>Value</th></tr>
<tr>
<td>fs.swift.impl</td>
<td>org.apache.hadoop.fs.swift.snative.SwiftNativeFileSystem</td>
</tr>
</table>
Additional parameters required by Keystone (v2.0) and should be provided to the Swift driver. Those
parameters will be used to perform authentication in Keystone to access Swift. The following table
contains a list of Keystone mandatory parameters. <code>PROVIDER</code> can be any name.
<table class="table">
<tr><th>Property Name</th><th>Meaning</th><th>Required</th></tr>
<tr>
<td><code>fs.swift.service.PROVIDER.auth.url</code></td>
<td>Keystone Authentication URL</td>
<td>Mandatory</td>
</tr>
<tr>
<td><code>fs.swift.service.PROVIDER.auth.endpoint.prefix</code></td>
<td>Keystone endpoints prefix</td>
<td>Optional</td>
</tr>
<tr>
<td><code>fs.swift.service.PROVIDER.tenant</code></td>
<td>Tenant</td>
<td>Mandatory</td>
</tr>
<tr>
<td><code>fs.swift.service.PROVIDER.username</code></td>
<td>Username</td>
<td>Mandatory</td>
</tr>
<tr>
<td><code>fs.swift.service.PROVIDER.password</code></td>
<td>Password</td>
<td>Mandatory</td>
</tr>
<tr>
<td><code>fs.swift.service.PROVIDER.http.port</code></td>
<td>HTTP port</td>
<td>Mandatory</td>
</tr>
<tr>
<td><code>fs.swift.service.PROVIDER.region</code></td>
<td>Keystone region</td>
<td>Mandatory</td>
</tr>
<tr>
<td><code>fs.swift.service.PROVIDER.public</code></td>
<td>Indicates if all URLs are public</td>
<td>Mandatory</td>
</tr>
</table>
For example, assume <code>PROVIDER=SparkTest</code> and Keystone contains user <code>tester</code> with password <code>testing</code>
defined for tenant <code>test</code>. Then <code>core-site.xml</code> should include:
{% highlight xml %}
<configuration>
<property>
<name>fs.swift.impl</name>
<value>org.apache.hadoop.fs.swift.snative.SwiftNativeFileSystem</value>
</property>
<property>
<name>fs.swift.service.SparkTest.auth.url</name>
<value>http://127.0.0.1:5000/v2.0/tokens</value>
</property>
<property>
<name>fs.swift.service.SparkTest.auth.endpoint.prefix</name>
<value>endpoints</value>
</property>
<name>fs.swift.service.SparkTest.http.port</name>
<value>8080</value>
</property>
<property>
<name>fs.swift.service.SparkTest.region</name>
<value>RegionOne</value>
</property>
<property>
<name>fs.swift.service.SparkTest.public</name>
<value>true</value>
</property>
<property>
<name>fs.swift.service.SparkTest.tenant</name>
<value>test</value>
</property>
<property>
<name>fs.swift.service.SparkTest.username</name>
<value>tester</value>
</property>
<property>
<name>fs.swift.service.SparkTest.password</name>
<value>testing</value>
</property>
</configuration>
{% endhighlight %}
Notice that
<code>fs.swift.service.PROVIDER.tenant</code>,
<code>fs.swift.service.PROVIDER.username</code>,
<code>fs.swift.service.PROVIDER.password</code> contains sensitive information and keeping them in
<code>core-site.xml</code> is not always a good approach.
We suggest to keep those parameters in <code>core-site.xml</code> for testing purposes when running Spark
via <code>spark-shell</code>.
For job submissions they should be provided via <code>sparkContext.hadoopConfiguration</code>.
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...