The hdfs2 component enables you to read and write messages from/to an HDFS file system using Hadoop 2.x. HDFS is the distributed file system at the heart of Hadoop.
Maven users will need to add the following dependency to their pom.xml
for
this component:
<dependency> <groupId>org.apache.camel</groupId> <artifactId>camel-hdfs2</artifactId> <version>x.x.x</version> <!-- use the same version as your Camel core version --> </dependency>
hdfs2://hostname[:port][/path][?options]
You can append query options to the URI in the following format,
?option=value&option=value&...
The path is treated in the following way:
as a consumer, if it's a file, it just reads the file, otherwise if it represents a directory it scans all the file under the path satisfying the configured pattern. All the files under that directory must be of the same type.
as a producer, if at least one split strategy is defined, the path is considered a directory and under that directory the producer creates a different file per split named using the configured UuidGenerator.
Name |
Default Value |
Description |
---|---|---|
|
|
The file can be overwritten |
|
|
Append to existing file. Notice that not all HDFS file systems support the append option. |
|
|
The buffer size used by HDFS |
|
|
The HDFS replication factor |
|
|
The size of the HDFS blocks |
|
|
It can be SEQUENCE_FILE, MAP_FILE, ARRAY_FILE, or BLOOMMAP_FILE, see Hadoop |
|
|
It can be LOCAL for local filesystem |
|
|
The type for the key in case of sequence or map files. See below. |
|
|
The type for the key in case of sequence or map files. See below. |
|
A string describing the strategy on how to split the file based on different criteria. See below. | |
|
|
When a file is opened for reading/writing the file is renamed with this suffix to avoid to read it during the writing phase. |
|
|
Once the file has been read is renamed with this suffix to avoid to read it again. |
|
|
For the consumer, how much to wait (milliseconds) before to start scanning the directory. |
|
|
The interval (milliseconds) between the directory scans. |
|
|
The pattern used for scanning the directory |
|
|
When reading a normal file, this is split into chunks producing a message per chunk. |
|
|
Camel 2.9.3/2.10.1: Whether to connect to the
HDFS file system on starting the producer/consumer. If |
|
Camel 2.13/2.12.4: The file owner must match this owner for the consumer to pickup the file. Otherwise the file is skipped. |
NULL it means that the key or the value is absent
BYTE for writing a byte, the java Byte class is mapped into a BYTE
BYTES for writing a sequence of bytes. It maps the java ByteBuffer class
INT for writing java integer
FLOAT for writing java float
LONG for writing java long
DOUBLE for writing java double
TEXT for writing java strings
BYTES is also used with everything else, for example, in Camel a file is sent around as an InputStream, int this case is written in a sequence file or a map file as a sequence of bytes.
In the current version of Hadoop opening a file in append mode is disabled since it's not very reliable. So, for the moment, it's only possible to create new files. The Camel HDFS endpoint tries to solve this problem in this way:
If the split strategy option has been defined, the hdfs path will be used as a directory and files will be created using the configured UuidGenerator
Every time a splitting condition is met, a new file is created.
The splitStrategy option is defined as a string with the following syntax:
splitStrategy=<ST>:<value>,<ST>:<value>,*
where <ST> can be:
BYTES a new file is created, and the old is closed when the number of written bytes is more than <value>
MESSAGES a new file is created, and the old is closed when the number of written messages is more than <value>
IDLE a new file is created, and the old is closed when no writing happened in the last <value> milliseconds
Note that this strategy currently requires either setting an IDLE value or setting the HdfsConstants.HDFS_CLOSE header to false to use the BYTES/MESSAGES configuration...otherwise, the file will be closed with each message
for example:
hdfs2://localhost/tmp/simple-file?splitStrategy=IDLE:1000,BYTES:5
it means: a new file is created either when it has been idle for more than 1 second or
if more than 5 bytes have been written. So, running hadoop fs -ls
/tmp/simple-file
you'll see that multiple files have been created.
The following headers are supported by this component:
Header |
Description |
---|---|
|
Camel 2.13: Specifies the name of the file to write (relative to the endpoint path). The name can be a String or an Expression object. Only relevant when not using a split strategy. |
When using the HDFS2
producer without a split strategy, then the file output
stream is by default closed after the write. However you may want to keep the stream open,
and only explicitly close the stream later. For that you can use the header
HdfsConstants.HDFS_CLOSE
(value = "CamelHdfsClose
") to control
this. Setting this value to a boolean allows you to explicit control whether the stream
should be closed or not.
Notice this does not apply if you use a split strategy, as there are various strategies that can control when the stream is closed.
There are some quirks when running this component in an OSGi environment related to the
mechanism Hadoop 2.x uses to discover different org.apache.hadoop.fs.FileSystem
implementations. Hadoop 2.x uses java.util.ServiceLoader
which looks for
/META-INF/services/org.apache.hadoop.fs.FileSystem
files defining available
filesystem types and implementations. These resources are not available when running inside
OSGi.
As with camel-hdfs
component, the default configuration files need to be
visible from the bundle class loader. A typical way to deal with it is to keep a copy of
core-default.xml
(and e.g., hdfs-default.xml
) in your bundle
root.
There are two options:
Package
/META-INF/services/org.apache.hadoop.fs.FileSystem
resource with bundle that defines the routes. This resource should list all the required Hadoop 2.x filesystem implementations.Provide boilerplate initialization code which populates internal, static cache inside
org.apache.hadoop.fs.FileSystem
class:org.apache.hadoop.conf.Configuration conf = new org.apache.hadoop.conf.Configuration(); conf.setClass("fs.file.impl", org.apache.hadoop.fs.LocalFileSystem.class, FileSystem.class); conf.setClass("fs.hdfs.impl", org.apache.hadoop.hdfs.DistributedFileSystem.class, FileSystem.class); ... FileSystem.get("file:///", conf); FileSystem.get("hdfs://localhost:9000/", conf); ...
Two options:
Package
/META-INF/services/org.apache.hadoop.fs.FileSystem
resource with bundle that contains blueprint definition.Add the following to the blueprint definition file:
<bean id="hdfsOsgiHelper" class="org.apache.camel.component.hdfs2.HdfsOsgiHelper"> <argument> <map> <entry key="file:///" value="org.apache.hadoop.fs.LocalFileSystem" /> <entry key="hdfs://localhost:9000/" value="org.apache.hadoop.hdfs.DistributedFileSystem" /> ... </map> </argument> </bean> <bean id="hdfs2" class="org.apache.camel.component.hdfs2.HdfsComponent" depends-on="hdfsOsgiHelper" />
This way Hadoop 2.x will have correct mapping of URI schemes to filesystem implementations.