Files
spring-integration/src/reference/docbook/ftp.xml
Artem Bilan 417c848c0b INT-3129: FTP local-filename-generator-expression
* Add `local-filename-generator-expression` attribute to FTP and SFTP Outbound Gateway
* Refactoring for `(S)FtpOutboundGatewayParser`
* Add `local-filename-generator-expression` tests
* Add 'What's New' and attribute description

JIRA: https://jira.springsource.org/browse/INT-3129

Polishing

Doc Polishing

Assert attribute is only set for get/mget

Move tests to the get gateways
2013-09-18 13:40:59 -04:00

498 lines
28 KiB
XML

<?xml version="1.0" encoding="UTF-8"?>
<chapter xmlns="http://docbook.org/ns/docbook" version="5.0" xml:id="ftp"
xmlns:xlink="http://www.w3.org/1999/xlink">
<title>FTP/FTPS Adapters</title>
<para>
Spring Integration provides support for file transfer operations via FTP and FTPS.
</para>
<section id="ftp-intro">
<title>Introduction</title>
<para>
The File Transfer Protocol (FTP) is a simple network protocol which allows you to transfer files between two computers on the Internet.
</para>
<para>
There are two actors when it comes to FTP communication: <emphasis>client</emphasis> and <emphasis>server</emphasis>.
To transfer files with FTP/FTPS, you use a <emphasis>client</emphasis> which initiates a connection to a remote computer
that is running an FTP <emphasis>server</emphasis>. After the connection is established, the <emphasis>client</emphasis> can choose
to send and/or receive copies of files.
</para>
<para>
Spring Integration supports sending and receiving files over FTP/FTPS by providing three <emphasis>client</emphasis>
side endpoints: <emphasis>Inbound Channel Adapter</emphasis>, <emphasis>Outbound Channel Adapter</emphasis>, and
<emphasis>Outbound Gateway</emphasis>. It also provides
convenient namespace-based configuration options for defining these <emphasis>client</emphasis> components.
</para>
<para>
To use the <emphasis>FTP</emphasis> namespace, add the following to the header of your XML file:
<programlisting language="xml"><![CDATA[xmlns:int-ftp="http://www.springframework.org/schema/integration/ftp"
xsi:schemaLocation="http://www.springframework.org/schema/integration/ftp
http://www.springframework.org/schema/integration/ftp/spring-integration-ftp.xsd"
]]></programlisting>
</para>
</section>
<section id="ftp-session-factory">
<title>FTP Session Factory</title>
<important>
Starting with version 3.0, sessions are no longer cached by default. See <xref linkend="ftp-session-caching"/>.
</important>
<para>
Before configuring FTP adapters you must configure an <emphasis>FTP Session Factory</emphasis>. You can configure
the <emphasis>FTP Session Factory</emphasis> with a regular bean definition where the implementation class is <classname>org.springframework.integration.ftp.session.DefaultFtpSessionFactory</classname>:
Below is a basic configuration:
<programlisting language="xml"><![CDATA[<bean id="ftpClientFactory"
class="org.springframework.integration.ftp.session.DefaultFtpSessionFactory">
<property name="host" value="localhost"/>
<property name="port" value="22"/>
<property name="username" value="kermit"/>
<property name="password" value="frog"/>
<property name="clientMode" value="0"/>
<property name="fileType" value="2"/>
<property name="bufferSize" value="100000"/>
</bean>]]></programlisting>
</para>
<para>
For FTPS connections all you need to do is use <classname>org.springframework.integration.ftp.session.DefaultFtpsSessionFactory</classname> instead.
Below is the complete configuration sample:
<programlisting language="xml"><![CDATA[<bean id="ftpClientFactory"
class="org.springframework.integration.ftp.client.DefaultFtpsClientFactory">
<property name="host" value="localhost"/>
<property name="port" value="22"/>
<property name="username" value="oleg"/>
<property name="password" value="password"/>
<property name="clientMode" value="1"/>
<property name="fileType" value="2"/>
<property name="useClientMode" value="true"/>
<property name="cipherSuites" value="a,b.c"/>
<property name="keyManager" ref="keyManager"/>
<property name="protocol" value="SSL"/>
<property name="trustManager" ref="trustManager"/>
<property name="prot" value="P"/>
<property name="needClientAuth" value="true"/>
<property name="authValue" value="oleg"/>
<property name="sessionCreation" value="true"/>
<property name="protocols" value="SSL, TLS"/>
<property name="implicit" value="true"/>
</bean>]]></programlisting>
</para>
<para>
Every time an adapter requests a session object from its <interfacename>SessionFactory</interfacename> the session is
returned from a session pool maintained by a caching wrapper around the factory. A Session in the session pool might go stale
(if it has been disconnected by the server due to inactivity) so the <interfacename>SessionFactory</interfacename>
will perform validation to make sure that it never returns a stale session to the adapter. If a stale session was encountered,
it will be removed from the pool, and a new one will be created.
<note>
If you experience connectivity problems and would like to trace Session creation as well as see which Sessions are
polled you may enable it by setting the logger to TRACE level (e.g., log4j.category.org.springframework.integration.file=TRACE)
</note>
</para>
<para>
Now all you need to do is inject these session factories into your adapters. Obviously the protocol (FTP or FTPS) that an adapter will
use depends on the type of session factory that has been injected into the adapter.
</para>
<para>
<note>
A more practical way to provide values for <emphasis>FTP/FTPS Session Factories</emphasis> is by using Spring's property
placeholder support (See: http://static.springsource.org/spring/docs/3.0.x/spring-framework-reference/html/beans.html#beans-factory-placeholderconfigurer).
</note>
</para>
<para>
<emphasis>Advanced Configuration</emphasis>
</para>
<para>
<classname>DefaultFtpSessionFactory</classname> provides an abstraction over the underlying client API which, in the current release of
Spring Integration, is <ulink url="http://commons.apache.org/net/">Apache Commons Net</ulink>. This spares you from the low level configuration details
of the <classname>org.apache.commons.net.ftp.FTPClient</classname>. However there are times when access to lower level <classname>FTPClient</classname> details is
necessary to achieve more advanced configuration (e.g., setting data timeout, default timeout etc.). For that purpose, <classname>AbstractFtpSessionFactory</classname>
(the base class for all FTP Session Factories) exposes hooks, in the form of the two post-processing methods below.
<programlisting language="java"><![CDATA[/**
* Will handle additional initialization after client.connect() method was invoked,
* but before any action on the client has been taken
*/
protected void postProcessClientAfterConnect(T t) throws IOException {
// NOOP
}
/**
* Will handle additional initialization before client.connect() method was invoked.
*/
protected void postProcessClientBeforeConnect(T client) throws IOException {
// NOOP
}]]></programlisting>
As you can see, there is no default implementation for these two methods. However, by extending <classname>DefaultFtpSessionFactory</classname> you can override these methods
to provide more advanced configuration of the <classname>FTPClient</classname>. For example:
<programlisting language="java"><![CDATA[public class AdvancedFtpSessionFactory extends DefaultFtpSessionFactory {
protected void postProcessClientBeforeConnect(FTPClient ftpClient) throws IOException {
ftpClient.setDataTimeout(5000);
ftpClient.setDefaultTimeout(5000);
}
}]]></programlisting>
</para>
</section>
<section id="ftp-inbound">
<title>FTP Inbound Channel Adapter</title>
<para>
The <emphasis>FTP Inbound Channel Adapter</emphasis> is a special listener that will connect to the FTP server and will listen
for the remote directory events (e.g., new file created) at which point it will initiate a file transfer.
<programlisting language="xml"><![CDATA[<int-ftp:inbound-channel-adapter id="ftpInbound"
channel="ftpChannel"
session-factory="ftpSessionFactory"
charset="UTF-8"
auto-create-local-directory="true"
delete-remote-files="true"
filename-pattern="*.txt"
remote-directory="some/remote/path"
remote-file-separator="/"
local-filename-generator-expression="#this.toUpperCase() + '.a'"
local-filter="myFilter"
local-directory=".">
<int:poller fixed-rate="1000"/>
</int-ftp:inbound-channel-adapter>]]></programlisting>
As you can see from the configuration above you can configure an <emphasis>FTP Inbound Channel Adapter</emphasis> via the <code>inbound-channel-adapter</code>
element while also providing values for various attributes such as <code>local-directory</code>, <code>filename-pattern</code>
(which is based on simple pattern matching, not regular expressions), and of course the reference to a <code>session-factory</code>.
</para>
<para>
By default the transferred file will carry the same name as the original file. If you want to override this behavior you
can set the <code>local-filename-generator-expression</code> attribute which allows you to provide a SpEL Expression to generate
the name of the local file. Unlike outbound gateways and adapters where the root object of the SpEL Evaluation Context
is a <classname>Message</classname>, this inbound adapter does not yet have the Message at the time of evaluation since
that's what it ultimately generates with the transferred file as its payload. So, the root object of the SpEL Evaluation Context
is the original name of the remote file (String).
</para>
<para>
Sometimes file filtering based on the simple pattern specified via <code>filename-pattern</code> attribute might not be
sufficient. If this is the case, you can use the <code>filename-regex</code> attribute to specify a Regular Expression
(e.g. <code>filename-regex=".*\.test$"</code>). And of course if you need complete control you can use <code>filter</code>
attribute and provide a reference to any custom implementation of the
<classname>org.springframework.integration.file.filters.FileListFilter</classname>, a strategy interface for filtering a
list of files. This filter determines which remote files are retrieved.
</para>
<note>
Beginning with 3.0, you can also specify a filter used to filter the files locally, once they have
been retrieved. The default filter is an <classname>AcceptOnceFileListFilter</classname> which prevents processing
files with the same name multiple times in the same JVM execution; this can now be overridden
(for example with an <classname>AcceptAllFileListFilter</classname>), using the <code>local-filter</code> attribute.
Previously, the default <classname>AcceptOnceFileListFilter</classname> could not be overridden.
</note>
<para>
The 'remote-file-separator' attribute allows you to configure a
file separator character to use if the default '/' is not applicable for your particular environment.
</para>
<para>
Please refer to the schema for more details on these attributes.
</para>
<para>
It is also important to understand that the <emphasis>FTP Inbound Channel Adapter</emphasis> is a <emphasis>Polling Consumer</emphasis> and
therefore you must configure a poller (either via a global default or a local sub-element).
Once a file has been transferred, a Message with a <classname>java.io.File</classname> as its payload will be generated and sent to the channel
identified by the <code>channel</code> attribute.
</para>
<para>
<emphasis>More on File Filtering and Large Files</emphasis>
</para>
<para>
Sometimes the file that just appeared in the monitored (remote) directory is not complete. Typically such a file
will be written with temporary extension (e.g., foo.txt.writing) and then renamed after the writing process finished.
As a user in most cases you are only interested in files that are complete and would like to filter only files that are complete.
To handle these scenarios you can use the filtering support provided by the <code>filename-pattern</code>, <code>filename-regex</code>
and <code>filter</code> attributes. Here is an example that uses a custom Filter implementation.
<programlisting language="xml"><![CDATA[<int-ftp:inbound-channel-adapter
channel="ftpChannel"
session-factory="ftpSessionFactory"
filter="customFilter"
local-directory="file:/my_transfers">
remote-directory="some/remote/path"
<int:poller fixed-rate="1000"/>
</int-ftp:inbound-channel-adapter>
<bean id="customFilter" class="org.example.CustomFilter"/>]]></programlisting>
</para>
<para>
<emphasis>Poller configuration notes for the inbound FTP adapter</emphasis>
</para>
<para>
The job of the inbound FTP adapter consists of two tasks:
<emphasis>1) Communicate with a remote server in order to transfer files from a remote directory to a local directory.</emphasis>
<emphasis>2) For each transferred file, generate a Message with that file as a payload and send it to the channel identified by the 'channel' attribute.</emphasis>
That is why they are called 'channel-adapters' rather than just 'adapters'. The main job of such an adapter is to generate a
Message to be sent to a Message Channel. Essentially, the second task mentioned above takes precedence in such a way that
*IF* your local directory already has one or more files it will first generate Messages from those, and *ONLY*
when all local files have been processed, will it initiate the remote communication to retrieve more files.
</para>
<para>
Also, when configuring a trigger on the poller you should pay close attention to the <code>max-messages-per-poll</code>
attribute. Its default value is 1 for all <classname>SourcePollingChannelAdapter</classname> instances (including FTP).
This means that as soon as one file is processed, it will wait for the next execution time as determined by your
trigger configuration. If you happened to have one or more files sitting in the <code>local-directory</code>, it would process
those files before it would initiate communication with the remote FTP server. And, if the <code>max-messages-per-poll</code>
were set to 1 (default), then it would be processing only one file at a time with intervals as defined by your trigger,
essentially working as <emphasis>one-poll = one-file</emphasis>.
</para>
<para>
For typical file-transfer use cases, you most likely want the opposite behavior: to process all the files you can for each
poll and only then wait for the next poll. If that is the case, set <code>max-messages-per-poll</code> to -1. Then, on
each poll, the adapter will attempt to generate as many Messages as it possibly can. In other words, it will process
everything in the local directory, and then it will connect to the remote directory to transfer everything that is available
there to be processed locally. Only then is the poll operation considered complete, and the poller will wait for the next execution time.
</para>
<para>
You can alternatively set the 'max-messages-per-poll' value to a positive value indicating the upward limit of Messages to be created
from files with each poll. For example, a value of 10 means that on each poll it will attempt to process no more than 10 files.
</para>
</section>
<section id="ftp-outbound">
<title>FTP Outbound Channel Adapter</title>
<para>
The <emphasis>FTP Outbound Channel Adapter</emphasis> relies upon a <classname>MessageHandler</classname> implementation that will connect to the
FTP server and initiate an FTP transfer for every file it receives in the payload of incoming Messages. It also supports several
representations of the <emphasis>File</emphasis> so you are not limited only to java.io.File typed payloads.
The <emphasis>FTP Outbound Channel Adapter</emphasis>
supports the following payloads: 1) <classname>java.io.File</classname> - the actual file object;
2) <classname>byte[]</classname> - a byte array that represents the file contents; and 3) <classname>java.lang.String</classname> -
text that represents the file contents.
<programlisting language="xml"><![CDATA[<int-ftp:outbound-channel-adapter id="ftpOutbound"
channel="ftpChannel"
session-factory="ftpSessionFactory"
charset="UTF-8"
remote-file-separator="/"
auto-create-directory="true"
remote-directory-expression="headers.['remote_dir']"
temporary-remote-directory-expression="headers.['temp_remote_dir']"
filename-generator="fileNameGenerator"/>]]></programlisting>
As you can see from the configuration above you can configure an <emphasis>FTP Outbound Channel Adapter</emphasis> via the
<code>outbound-channel-adapter</code> element while also providing values for various attributes such as <code>filename-generator</code>
(an implementation of the <classname>org.springframework.integration.file.FileNameGenerator</classname> strategy interface),
a reference to a <code>session-factory</code>, as well as other attributes. You can also see
some examples of <code>*expression</code> attributes which allow you to use SpEL
to configure things like <code>remote-directory-expression</code>, <code>temporary-remote-directory-expression</code> and <code>remote-filename-generator-expression</code>
(a SpEL alternative to <code>filename-generator</code> shown above). As with any component that allows the usage of SpEL, access to Payload and Message Headers is available via
'payload' and 'headers' variables.
Please refer to the schema for more details on
the available attributes.
<note>
By default Spring Integration will use <classname>o.s.i.file.DefaultFileNameGenerator</classname> if none is specified.
<classname>DefaultFileNameGenerator</classname> will determine the file name based on the value of the <code>file_name</code> header (if it exists)
in the MessageHeaders, or if the payload of the Message is already a <classname>java.io.File</classname>, then it will use the original name of that file.
</note>
</para>
<para>
<important>
Defining certain values (e.g., remote-directory) might be platform/ftp server dependent. For example as it
was reported on this forum http://forum.springsource.org/showthread.php?p=333478&amp;posted=1#post333478 on some
platforms you must add slash to the end of the directory definition (e.g., remote-directory="/foo/bar/"
instead of remote-directory="/foo/bar")
</important>
</para>
<para>
<emphasis>Avoiding Partially Written Files</emphasis>
</para>
<para>
One of the common problems, when dealing with file transfers, is the possibility of processing a <emphasis>partial file</emphasis> -
a file might appear in the file system before its transfer is actually complete.
</para>
<para>
To deal with this issue, Spring Integration FTP adapters use a very common algorithm where files are transferred
under a temporary name and then renamed once they are fully transferred.
</para>
<para>
By default, every file that is in the process of being transferred will appear in the file system with an additional suffix
which, by default, is <code>.writing</code>; this can be changed using the <code>temporary-file-suffix</code> attribute.
</para>
<para>
However, there may be situations where you don't want to use this technique (for example, if the server does not
permit renaming files). For situations like this, you can disable this feature by setting <code>use-temporary-file-name</code>
to <code>false</code> (default is <code>true</code>). When this attribute is <code>false</code>, the file is written with its
final name and the consuming application will need some other mechanism to detect that the file is completely uploaded before accessing it.
</para>
</section>
<section id="ftp-outbound-gateway">
<title>FTP Outbound Gateway</title>
<para>
The <emphasis>FTP Outbound Gateway</emphasis> provides a limited set of commands to interact with a remote FTP/FTPS server.
<para>
Commands supported are:
<itemizedlist>
<listitem>ls (list files)</listitem>
<listitem>get (retrieve file)</listitem>
<listitem>mget (retrieve file(s))</listitem>
<listitem>rm (remove file(s))</listitem>
<listitem>mv (move/rename file)</listitem>
</itemizedlist>
</para>
<para><emphasis role="bold">ls</emphasis></para>
<para>
ls lists remote file(s) and supports the following options:
<itemizedlist>
<listitem>-1 - just retrieve a list of filenames, default is to retrieve a
list of <classname>FileInfo</classname> objects.</listitem>
<listitem>-a - include all files (including those starting with '.')</listitem>
<listitem>-f - do not sort the list</listitem>
<listitem>-dirs - include directories (excluded by default)</listitem>
<listitem>-links - include symbolic links (excluded by default)</listitem>
</itemizedlist>
</para>
<para>
In addition, filename filtering is provided, in the same manner as the
<classname>inbound-channel-adapter</classname>.
</para>
<para>
The message payload resulting from an <emphasis>ls</emphasis> operation is a list of file names,
or a list of <classname>FileInfo</classname> objects. These objects provide
information such as modified time, permissions etc.
</para>
<para>
The remote directory that the <emphasis>ls</emphasis> command acted on is provided
in the <classname>file_remoteDirectory</classname> header.
</para>
<para><emphasis role="bold">get</emphasis></para>
<para>
<emphasis>get</emphasis> retrieves a remote file and supports the following option:
<itemizedlist>
<listitem>-P - preserve the timestamp of the remote file</listitem>
</itemizedlist>
</para>
<para>
The message payload resulting from a <emphasis>get</emphasis> operation is a
<classname>File</classname> object representing the retrieved file.
</para>
<para>
The remote directory is provided in the <classname>file_remoteDirectory</classname> header, and the filename is
provided in the <classname>file_remoteFile</classname> header.
</para>
<para><emphasis role="bold">mget</emphasis></para>
<para>
<emphasis>mget</emphasis> retrieves multiple remote files based on a pattern and supports the following option:
<itemizedlist>
<listitem>-x - Throw an exception if no files match the pattern (otherwise an empty
list is returned)</listitem>
</itemizedlist>
</para>
<para>
The message payload resulting from an <emphasis>mget</emphasis> operation is a
<classname>List&lt;File&gt;</classname> object - a List of File objects, each representing
a retrieved file.
</para>
<para>
The remote directory is provided in the <classname>file_remoteDirectory</classname> header, and the pattern
for the filenames is
provided in the <classname>file_remoteFile</classname> header.
</para>
<para><emphasis role="bold">rm</emphasis></para>
<para>
The <emphasis>rm</emphasis> command has no options.
</para>
<para>
The message payload resulting from an <emphasis>rm</emphasis> operation is Boolean.TRUE if the
remove was successful, Boolean.FALSE otherwise.
The remote directory is provided in the <classname>file_remoteDirectory</classname> header, and the filename is
provided in the <classname>file_remoteFile</classname> header.
</para>
<para><emphasis role="bold">mv</emphasis></para>
<para>
The <emphasis>mv</emphasis> command has no options.
</para>
<para>
The <emphasis>expression</emphasis> attribute defines the "from" path and the
<emphasis>rename-expression</emphasis> attribute defines the "to" path. By default, the
<emphasis>rename-expression</emphasis> is <code>headers['file_renameTo']</code>. This
expression must not evaluate to null, or an empty <code>String</code>. If necessary,
any remote directories needed will be created.
The payload of the result message is <code>Boolean.TRUE</code>.
The original remote directory is provided in the <code>file_remoteDirectory</code> header, and the filename is
provided in the <code>file_remoteFile</code> header. The new path is in
the <code>file_renameTo</code> header.
</para>
<para>
<emphasis role="bold">Additional Information</emphasis>
</para>
<para>
The <emphasis>get</emphasis> and <emphasis>mget</emphasis> commands support
the <emphasis>local-filename-generator-expression</emphasis> attribute. It
defines a SpEL expression to generate the name of local file(s) during the transfer.
The root object of the evaluation context is the request Message but, in addition, the <code>remoteFileName</code>
variable is also available, which is particularly useful for <emphasis>mget</emphasis>, for
example: <code>local-filename-generator-expression="#remoteFileName.toUpperCase() + headers.foo"</code>
</para>
<para>
For all commands, the PATH that the command acts on is provided by the 'expression'
property of the gateway. For the mget command, the expression might evaluate to '*', meaning
retrieve all files, or 'somedirectory/*' etc.
</para>
</para>
<para>
Here is an example of a gateway configured for an ls command...
<programlisting language="xml"><![CDATA[<int-ftp:outbound-gateway id="gateway1"
session-factory="ftpSessionFactory"
request-channel="inbound1"
command="ls"
command-options="-1"
expression="payload"
reply-channel="toSplitter"/>]]></programlisting>
</para>
<para>
The payload of the message sent to the toSplitter channel is a list of String objects
containing the filename of each file. If the <classname>command-options</classname> was
omitted, it would be a list of <classname>FileInfo</classname> objects. Options are
provided space-delimited, e.g. <classname>command-options="-1 -dirs -links"</classname>.
</para>
</section>
<section id="ftp-session-caching">
<title>FTP Session Caching</title>
<important>
Starting with version 3.0, sessions are no longer cached by default; the <code>cache-sessions</code> attribute
is no longer supported on endpoints. You must now use a <classname>CachingSessionFactory</classname> (see below) if you
wish to cache sessions.
</important>
<para>
In versions prior to 3.0, the sessions were cached automatically by default. A <code>cache-sessions</code> attribute was available for
disabling the auto caching, but that solution did not provide a way to configure other session caching attributes. For example,
you could not limit on the number of sessions created. To support that requirement and other configuration options, a
<classname>CachingSessionFactory</classname> was provided. It provides <code>sessionCacheSize</code> and <code>sessionWaitTimeout</code>
properties. As its name suggests, the <code>sessionCacheSize</code> property controls how many active sessions the factory will
maintain in its cache (the DEFAULT is unbounded). If the <code>sessionCacheSize</code> threshold has been reached, any attempt to
acquire another session will block until either one of the cached sessions becomes available or until the wait time for a Session
expires (the DEFAULT wait time is Integer.MAX_VALUE). The <code>sessionWaitTimeout</code> property enables configuration of that value.
</para>
<para>
If you want your Sessions to be cached, simply configure your default Session Factory as described above and then
wrap it in an instance of <classname>CachingSessionFactory</classname> where you may provide those additional properties.
</para>
<programlisting language="xml"><![CDATA[<bean id="ftpSessionFactory" class="o.s.i.ftp.session.DefaultFtpSessionFactory">
<property name="host" value="localhost"/>
</bean>
<bean id="cachingSessionFactory" class="o.s.i.file.remote.session.CachingSessionFactory">
<constructor-arg ref="ftpSessionFactory"/>
<constructor-arg value="10"/>
<property name="sessionWaitTimeout" value="1000"/>
</bean>]]></programlisting>
<para>
In the above example you see a <classname>CachingSessionFactory</classname> created with the
<code>sessionCacheSize</code> set to 10 and the <code>sessionWaitTimeout</code> set to 1 second (its value is in millliseconds).
</para>
</section>
</chapter>