true if the key is deprecated and
+ false otherwise.]]>
+ null if
+ no such property exists. If the key is deprecated, it returns the value of
+ the first key which replaces the deprecated key and is not null.
+
+ Values are processed for variable expansion
+ before being returned.
+
+ @param name the property name, will be trimmed before get value.
+ @return the value of the name or its replacing property,
+ or null if no such property exists.]]>
+ name exists without value]]>
+ String,
+ null if no such property exists.
+ If the key is deprecated, it returns the value of
+ the first key which replaces the deprecated key and is not null
+
+ Values are processed for variable expansion
+ before being returned.
+
+ @param name the property name.
+ @return the value of the name or its replacing property,
+ or null if no such property exists.]]>
+ String,
+ defaultValue if no such property exists.
+ See @{Configuration#getTrimmed} for more details.
+
+ @param name the property name.
+ @param defaultValue the property default value.
+ @return the value of the name or defaultValue
+ if it is not set.]]>
+ name property or
+ its replacing property and null if no such property exists.]]>
+ name property. If
+ name is deprecated or there is a deprecated name associated to it,
+ it sets the value to both names. Name will be trimmed before put into
+ configuration.
+
+ @param name property name.
+ @param value property value.]]>
+ name property. If
+ name is deprecated, it also sets the value to
+ the keys that replace the deprecated key. Name will be trimmed before put
+ into configuration.
+
+ @param name property name.
+ @param value property value.
+ @param source the place that this configuration value came from
+ (For debugging).
+ @throws IllegalArgumentException when the value or name is null.]]>
+ defaultValue is returned.
+
+ @param name property name, will be trimmed before get value.
+ @param defaultValue default value.
+ @return property value, or defaultValue if the property
+ doesn't exist.]]>
+ int.
+
+ If no such property exists, the provided default value is returned,
+ or if the specified value is not a valid int,
+ then an error is thrown.
+
+ @param name property name.
+ @param defaultValue default value.
+ @throws NumberFormatException when the value is invalid
+ @return property value as an int,
+ or defaultValue.]]>
+ int values.
+
+ If no such property exists, an empty array is returned.
+
+ @param name property name
+ @return property value interpreted as an array of comma-delimited
+ int values]]>
+ int.
+
+ @param name property name.
+ @param value int value of the property.]]>
+ long.
+ If no such property exists, the provided default value is returned,
+ or if the specified value is not a valid long,
+ then an error is thrown.
+
+ @param name property name.
+ @param defaultValue default value.
+ @throws NumberFormatException when the value is invalid
+ @return property value as a long,
+ or defaultValue.]]>
+ long or
+ human readable format. If no such property exists, the provided default
+ value is returned, or if the specified value is not a valid
+ long or human readable format, then an error is thrown. You
+ can use the following suffix (case insensitive): k(kilo), m(mega), g(giga),
+ t(tera), p(peta), e(exa)
+
+ @param name property name.
+ @param defaultValue default value.
+ @throws NumberFormatException when the value is invalid
+ @return property value as a long,
+ or defaultValue.]]>
+ long.
+
+ @param name property name.
+ @param value long value of the property.]]>
+ float.
+ If no such property exists, the provided default value is returned,
+ or if the specified value is not a valid float,
+ then an error is thrown.
+
+ @param name property name.
+ @param defaultValue default value.
+ @throws NumberFormatException when the value is invalid
+ @return property value as a float,
+ or defaultValue.]]>
+ float.
+
+ @param name property name.
+ @param value property value.]]>
+ double.
+ If no such property exists, the provided default value is returned,
+ or if the specified value is not a valid double,
+ then an error is thrown.
+
+ @param name property name.
+ @param defaultValue default value.
+ @throws NumberFormatException when the value is invalid
+ @return property value as a double,
+ or defaultValue.]]>
+ double.
+
+ @param name property name.
+ @param value property value.]]>
+ boolean.
+ If no such property is specified, or if the specified value is not a valid
+ boolean, then defaultValue is returned.
+
+ @param name property name.
+ @param defaultValue default value.
+ @return property value as a boolean,
+ or defaultValue.]]>
+ boolean.
+
+ @param name property name.
+ @param value boolean value of the property.]]>
+ set(<name>, value.toString()).
+ @param name property name
+ @param value new value]]>
+ set(<name>, value + <time suffix>).
+ @param name Property name
+ @param value Time duration
+ @param unit Unit of time]]>
+ Pattern.
+ If no such property is specified, or if the specified value is not a valid
+ Pattern, then DefaultValue is returned.
+ Note that the returned value is NOT trimmed by this method.
+
+ @param name property name
+ @param defaultValue default value
+ @return property value as a compiled Pattern, or defaultValue]]>
+ Strings.
+ If no such property is specified then empty collection is returned.
+
+ This is an optimized version of {@link #getStrings(String)}
+
+ @param name property name.
+ @return property value as a collection of Strings.]]>
+
Strings.
+ If no such property is specified then null is returned.
+
+ @param name property name.
+ @return property value as an array of Strings,
+ or null.]]>
+ Strings.
+ If no such property is specified then default value is returned.
+
+ @param name property name.
+ @param defaultValue The default value
+ @return property value as an array of Strings,
+ or default value.]]>
+ Strings, trimmed of the leading and trailing whitespace.
+ If no such property is specified then empty Collection is returned.
+
+ @param name property name.
+ @return property value as a collection of Strings, or empty Collection]]>
+ Strings, trimmed of the leading and trailing whitespace.
+ If no such property is specified then an empty array is returned.
+
+ @param name property name.
+ @return property value as an array of trimmed Strings,
+ or empty array.]]>
+ Strings, trimmed of the leading and trailing whitespace.
+ If no such property is specified then default value is returned.
+
+ @param name property name.
+ @param defaultValue The default value
+ @return property value as an array of trimmed Strings,
+ or default value.]]>
+ InetSocketAddress. If hostProperty is
+ null, addressProperty will be used. This
+ is useful for cases where we want to differentiate between host
+ bind address and address clients should use to establish connection.
+
+ @param hostProperty bind host property name.
+ @param addressProperty address property name.
+ @param defaultAddressValue the default value
+ @param defaultPort the default port
+ @return InetSocketAddress]]>
+ InetSocketAddress.
+ @param name property name.
+ @param defaultAddress the default value
+ @param defaultPort the default port
+ @return InetSocketAddress]]>
+ host:port.]]>
+ host:port. The wildcard
+ address is replaced with the local host's address. If the host and address
+ properties are configured the host component of the address will be combined
+ with the port component of the addr to generate the address. This is to allow
+ optional control over which host name is used in multi-home bind-host
+ cases where a host can have multiple names
+ @param hostProperty the bind-host configuration name
+ @param addressProperty the service address configuration name
+ @param defaultAddressValue the service default address configuration value
+ @param addr InetSocketAddress of the service listener
+ @return InetSocketAddress for clients to connect]]>
+ host:port. The wildcard
+ address is replaced with the local host's address.
+ @param name property name.
+ @param addr InetSocketAddress of a listener to store in the given property
+ @return InetSocketAddress for clients to connect]]>
+ Class.
+ The value of the property specifies a list of comma separated class names.
+ If no such property is specified, then defaultValue is
+ returned.
+
+ @param name the property name.
+ @param defaultValue default value.
+ @return property value as a Class[],
+ or defaultValue.]]>
+ Class.
+ If no such property is specified, then defaultValue is
+ returned.
+
+ @param name the class name.
+ @param defaultValue default value.
+ @return property value as a Class,
+ or defaultValue.]]>
+ Class
+ implementing the interface specified by xface.
+
+ If no such property is specified, then defaultValue is
+ returned.
+
+ An exception is thrown if the returned class does not implement the named
+ interface.
+
+ @param name the class name.
+ @param defaultValue default value.
+ @param xface the interface implemented by the named class.
+ @return property value as a Class,
+ or defaultValue.]]>
+ List
+ of objects implementing the interface specified by xface.
+
+ An exception is thrown if any of the classes does not exist, or if it does
+ not implement the named interface.
+
+ @param name the property name.
+ @param xface the interface implemented by the classes named by
+ name.
+ @return a List of objects implementing xface.]]>
+ theClass implementing the given interface xface.
+
+ An exception is thrown if theClass does not implement the
+ interface xface.
+
+ @param name property name.
+ @param theClass property value.
+ @param xface the interface implemented by the named class.]]>
+ + +
+ +
+ @param out the writer to write to.]]> +
+ {
+ "property": {
+ "key" : "key1",
+ "value" : "value1",
+ "isFinal" : "key1.isFinal",
+ "resource" : "key1.resource"
+ }
+ }
+
+
+
+
+ { "properties" :
+ [ { key : "key1",
+ value : "value1",
+ isFinal : "key1.isFinal",
+ resource : "key1.resource" },
+ { key : "key2",
+ value : "value2",
+ isFinal : "ke2.isFinal",
+ resource : "key2.resource" }
+ ]
+ }
+
+ + @param config the configuration + @param propertyName property name + @param out the Writer to write to + @throws IOException + @throws IllegalArgumentException when property name is not + empty and the property is not found in configuration]]> +
+ + @param config the configuration + @param out the Writer to write to + @throws IOException]]> +
false
+ to turn it off.]]>
+ Configurations are specified by resources. A resource contains a set of
+ name/value pairs as XML data. Each resource is named by either a
+ String or by a {@link Path}. If named by a String,
+ then the classpath is examined for a file with that name. If named by a
+ Path, then the local filesystem is examined directly, without
+ referring to the classpath.
+
+
Unless explicitly turned off, Hadoop by default specifies two + resources, loaded in-order from the classpath:
Configuration parameters may be declared final.
+ Once a resource declares a value final, no subsequently-loaded
+ resource can alter that value.
+ For example, one might define a final parameter with:
+
+ <property>
+ <name>dfs.hosts.include</name>
+ <value>/etc/hadoop/conf/hosts.include</value>
+ <final>true</final>
+ </property>
+
+ Administrators typically define parameters as final in
+ core-site.xml for values that user applications may not alter.
+
+
Value strings are first processed for variable expansion. The + available properties are:
For example, if a configuration resource contains the following property
+ definitions:
+
+ <property>
+ <name>basedir</name>
+ <value>/user/${user.name}</value>
+ </property>
+
+ <property>
+ <name>tempdir</name>
+ <value>${basedir}/tmp</value>
+ </property>
+
+ <property>
+ <name>otherdir</name>
+ <value>${env.BASE_DIR}/other</value>
+ </property>
+
+
+
When conf.get("tempdir") is called, then ${basedir} + will be resolved to another property in this Configuration, while + ${user.name} would then ordinarily be resolved to the value + of the System property with that name. +
When conf.get("otherdir") is called, then ${env.BASE_DIR} + will be resolved to the value of the ${BASE_DIR} environment variable. + It supports ${env.NAME:-default} and ${env.NAME-default} notations. + The former is resolved to "default" if ${NAME} environment variable is undefined + or its value is empty. + The latter behaves the same way only if ${NAME} is undefined. +
By default, warnings will be given to any deprecated configuration + parameters and these are suppressible by configuring + log4j.logger.org.apache.hadoop.conf.Configuration.deprecation in + log4j.properties file. + +
Optionally we can tag related properties together by using tag + attributes. System tags are defined by hadoop.system.tags property. Users + can define there own custom tags in hadoop.custom.tags property. + +
For example, we can tag existing property as:
+
+ <property>
+ <name>dfs.replication</name>
+ <value>3</value>
+ <tag>HDFS,REQUIRED</tag>
+ </property>
+
+ <property>
+ <name>dfs.data.transfer.protection</name>
+ <value>3</value>
+ <tag>HDFS,SECURITY</tag>
+ </property>
+
+
Properties marked with tags can be retrieved with conf + .getAllPropertiesByTag("HDFS") or conf.getAllPropertiesByTags + (Arrays.asList("YARN","SECURITY")).
]]> +KeyProvider implementations must be thread safe.]]>
+ uri is not supported.]]>
+
+ BlockLocation(offset: 0, length: 3 * BLOCK_SIZE, hosts: {"host1:9866",
+ "host2:9866","host3:9866","host4:9866","host5:9866"})
+
+
+ Please refer to
+ {@link FileSystem#getFileBlockLocations(FileStatus, long, long)} or
+ {@link FileContext#getFileBlockLocations(Path, long, long)}
+ for more examples.]]>
+
+ EnumSet.of(CreateFlag.CREATE, CreateFlag.APPEND)
+
+
+
+ Use the CreateFlag as follows:
+
+ - CREATE - to create a file if it does not exist,
+ else throw FileAlreadyExists.
+ - APPEND - to append to a file if it exists,
+ else throw FileNotFoundException.
+ - OVERWRITE - to truncate a file if it exists,
+ else throw FileNotFoundException.
+ - CREATE|APPEND - to create a file if it does not exist,
+ else append to an existing file.
+ - CREATE|OVERWRITE - to create a file if it does not exist,
+ else overwrite an existing file.
+ - SYNC_BLOCK - to force closed blocks to the disk device.
+ In addition {@link Syncable#hsync()} should be called after each write,
+ if true synchronous behavior is required.
+ - LAZY_PERSIST - Create the block on transient storage (RAM) if
+ available.
+ - APPEND_NEWBLOCK - Append data to a new block instead of end of the last
+ partial block.
+
+
+ Following combinations are not valid and will result in
+ {@link HadoopIllegalArgumentException}:
+
+ - APPEND|OVERWRITE
+ - CREATE|APPEND|OVERWRITE
+
]]>
+ absOrFqPath could
+ not be instantiated.]]>
+ f is not valid]]>
+ f already exists
+ @throws FileNotFoundException If parent of f does not exist
+ and createParent is false
+ @throws ParentNotDirectoryException If parent of f is not a
+ directory.
+ @throws UnsupportedFileSystemException If file system for f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server
+
+ RuntimeExceptions:
+ @throws InvalidPathException If path f is not valid]]>
+ dir does not exist
+ and createParent is false
+ @throws ParentNotDirectoryException If parent of dir is not a
+ directory
+ @throws UnsupportedFileSystemException If file system for dir
+ is not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server
+
+ RuntimeExceptions:
+ @throws InvalidPathException If path dir is not valid]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server
+
+ RuntimeExceptions:
+ @throws InvalidPathException If path f is invalid]]>
+ f
+ is not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ true if the file has been truncated to the desired
+ newLength and is immediately available to be reused for
+ write operations such as append, or
+ false if a background process of adjusting the length of
+ the last block has been started, and clients should wait for it to
+ complete before proceeding with further file updates.
+
+ @throws AccessControlException If access is denied
+ @throws FileNotFoundException If file f does not exist
+ @throws UnsupportedFileSystemException If file system for f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ + If OVERWRITE option is not passed as an argument, rename fails if the dst + already exists. +
+ If OVERWRITE option is passed as an argument, rename overwrites the dst if + it is a file or an empty directory. Rename fails if dst is a non-empty + directory. +
+ Note that atomicity of rename is dependent on the file system + implementation. Please refer to the file system documentation for details +
+
+ @param src path to be renamed
+ @param dst new path after rename
+
+ @throws AccessControlException If access is denied
+ @throws FileAlreadyExistsException If dst already exists and
+ options has {@link Options.Rename#OVERWRITE}
+ option false.
+ @throws FileNotFoundException If src does not exist
+ @throws ParentNotDirectoryException If parent of dst is not a
+ directory
+ @throws UnsupportedFileSystemException If file system for src
+ and dst is not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+
f
+ is not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server
+
+ RuntimeExceptions:
+ @throws HadoopIllegalArgumentException If username or
+ groupname is invalid.]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred]]>
+ f is
+ not supported
+ @throws IOException If the given path does not refer to a symlink
+ or an I/O error occurred]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ linkcode> already exists
+ @throws FileNotFoundException If target does not exist
+ @throws ParentNotDirectoryException If parent of link is not a
+ directory.
+ @throws UnsupportedFileSystemException If file system for
+ target or link is not supported
+ @throws IOException If an I/O error occurred]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ f is
+ not supported
+ @throws IOException If an I/O error occurred
+
+ Exceptions applicable to file systems accessed over RPC:
+ @throws RpcClientException If an exception occurred in the RPC client
+ @throws RpcServerException If an exception occurred in the RPC server
+ @throws UnexpectedServerException If server implementation throws
+ undeclared exception to RPC server]]>
+ + + Hadoop also supports working-directory-relative names, which are paths + relative to the current working directory (similar to Unix). The working + directory can be in a different file system than the default FS. +
+ Thus, Hadoop path names can be specified as one of the following: +
+ The file system related server-side defaults are: +
UnsupportedOperationException.
+
+ @return the protocol scheme for this FileSystem.
+ @throws UnsupportedOperationException if the operation is unsupported
+ (default).]]>
+
+ BlockLocation( { "localhost:9866" }, { "localhost" }, 0, file.getLen())
+
+
+ In HDFS, if file is three-replicated, the returned array contains
+ elements like:
+
+ BlockLocation(offset: 0, length: BLOCK_SIZE,
+ hosts: {"host1:9866", "host2:9866, host3:9866"})
+ BlockLocation(offset: BLOCK_SIZE, length: BLOCK_SIZE,
+ hosts: {"host2:9866", "host3:9866, host4:9866"})
+
+
+ And if a file is erasure-coded, the returned BlockLocation are logical
+ block groups.
+
+ Suppose we have a RS_3_2 coded file (3 data units and 2 parity units).
+ 1. If the file size is less than one stripe size, say 2 * CELL_SIZE, then
+ there will be one BlockLocation returned, with 0 offset, actual file size
+ and 4 hosts (2 data blocks and 2 parity blocks) hosting the actual blocks.
+ 3. If the file size is less than one group size but greater than one
+ stripe size, then there will be one BlockLocation returned, with 0 offset,
+ actual file size with 5 hosts (3 data blocks and 2 parity blocks) hosting
+ the actual blocks.
+ 4. If the file size is greater than one group size, 3 * BLOCK_SIZE + 123
+ for example, then the result will be like:
+
+ BlockLocation(offset: 0, length: 3 * BLOCK_SIZE, hosts: {"host1:9866",
+ "host2:9866","host3:9866","host4:9866","host5:9866"})
+ BlockLocation(offset: 3 * BLOCK_SIZE, length: 123, hosts: {"host1:9866",
+ "host4:9866", "host5:9866"})
+
+
+ @param file FilesStatus to get data from
+ @param start offset into the given file
+ @param len length for which to get locations for
+ @throws IOException IO failure]]>
+ + If OVERWRITE option is not passed as an argument, rename fails + if the dst already exists. +
+ If OVERWRITE option is passed as an argument, rename overwrites + the dst if it is a file or an empty directory. Rename fails if dst is + a non-empty directory. +
+ Note that atomicity of rename is dependent on the file system + implementation. Please refer to the file system documentation for + details. This default implementation is non atomic. +
+ This method is deprecated since it is a temporary method added to + support the transition from FileSystem to FileContext for user + applications. + + @param src path to be renamed + @param dst new path after rename + @throws FileNotFoundException src path does not exist, or the parent + path of dst does not exist. + @throws FileAlreadyExistsException dest path exists and is a file + @throws ParentNotDirectoryException if the parent path of dest is not + a directory + @throws IOException on failure]]> +
true if the file has been truncated to the desired
+ newLength and is immediately available to be reused for
+ write operations such as append, or
+ false if a background process of adjusting the length of
+ the last block has been started, and clients should wait for it to
+ complete before proceeding with further file updates.
+ @throws IOException IO failure
+ @throws UnsupportedOperationException if the operation is unsupported
+ (default).]]>
+ + Will not return null. Expect IOException upon access error. + @param f given path + @return the statuses of the files/directories in the given patch + @throws FileNotFoundException when the path does not exist + @throws IOException see specific implementation]]> +
+ A filename pattern is composed of regular characters and + special pattern matching characters, which are: + +
+
+
+
+
+
+
+
+
+ The local implementation is {@link LocalFileSystem} and distributed + implementation is DistributedFileSystem. There are other implementations + for object stores and (outside the Apache Hadoop codebase), + third party filesystems. +
+ Notes +
FilterFileSystem
+ itself simply overrides all methods of
+ FileSystem with versions that
+ pass all requests to the contained file
+ system. Subclasses of FilterFileSystem
+ may further override some of these methods
+ and may also provide additional methods
+ and fields.]]>
+ file]]>
+ pathname
+ should be included]]>
+ ftp]]>
+ + To use viewfs one would typically set the default file system in the + config (i.e. fs.defaultFS < = viewfs:///) along with the + mount table config variables as described below. + +
+ ** Config variables to specify the mount table entries ** +
+ + The file system is initialized from the standard Hadoop config through + config variables. + See {@link FsConstants} for URI and Scheme constants; + See {@link Constants} for config var constants; + see {@link ConfigUtil} for convenient lib. + +
+ All the mount table config entries for view fs are prefixed by + fs.viewfs.mounttable. + For example the above example can be specified with the following + config variables: +
+ **** Merge Mounts **** (NOTE: merge mounts are not implemented yet.) +
+ + One can also use "MergeMounts" to merge several directories (this is + sometimes called union-mounts or junction-mounts in the literature. + For example of the home directories are stored on say two file systems + (because they do not fit on one) then one could specify a mount + entry such as following merges two dirs: +
+ Fencing is configured by the operator as an ordered list of methods to + attempt. Each method will be tried in turn, and the next in the list + will only be attempted if the previous one fails. See {@link NodeFencer} + for more information. +
+ If an implementation also implements {@link Configurable} then its
+ setConf method will be called upon instantiation.]]>
+
+ Compared with ObjectWritable, this class is much more effective,
+ because ObjectWritable will append the class declaration as a String
+ into the output file in every Key-Value pair.
+
+ Generic Writable implements {@link Configurable} interface, so that it will be + configured by the framework. The configuration is passed to the wrapped objects + implementing {@link Configurable} interface before deserialization. +
+ + how to use it:getTypes(), defines
+ the classes which will be wrapped in GenericObject in application.
+ Attention: this classes defined in getTypes() method, must
+ implement Writable interface.
+
+ public class GenericObject extends GenericWritable {
+
+ private static Class[] CLASSES = {
+ ClassType1.class,
+ ClassType2.class,
+ ClassType3.class,
+ };
+
+ protected Class[] getTypes() {
+ return CLASSES;
+ }
+
+ }
+ data file,
+ containing all keys and values in the map, and a smaller index
+ file, containing a fraction of the keys. The fraction is determined by
+ {@link Writer#getIndexInterval()}.
+
+ The index file is read entirely into memory. Thus key implementations + should try to keep themselves small. + +
Map files are created by adding entries in-order. To maintain a large + database, perform updates by copying the previous version of a database and + merging in a sorted change list, to create a new version of the database in + a new file. Sorting large change lists can be done with {@link + SequenceFile.Sorter}.]]> +
SequenceFile provides {@link SequenceFile.Writer},
+ {@link SequenceFile.Reader} and {@link Sorter} classes for writing,
+ reading and sorting respectively.
SequenceFile Writers based on the
+ {@link CompressionType} used to compress key/value pairs:
+ Writer : Uncompressed records.
+ RecordCompressWriter : Record-compressed files, only compress
+ values.
+ BlockCompressWriter : Block-compressed files, both keys &
+ values are collected in 'blocks'
+ separately and compressed. The size of
+ the 'block' is configurable.
+ The actual compression algorithm used to compress key and/or values can be + specified by using the appropriate {@link CompressionCodec}.
+ +The recommended way is to use the static createWriter methods
+ provided by the SequenceFile to chose the preferred format.
The {@link SequenceFile.Reader} acts as the bridge and can read any of the
+ above SequenceFile formats.
Essentially there are 3 different formats for SequenceFiles
+ depending on the CompressionType specified. All of them share a
+ common header described below.
+
+
CompressionCodec class which is used for
+ compression of keys and/or values (if compression is
+ enabled).
+ 100 kilobytes or so.
+ 100 kilobytes or so.
+ The compressed blocks of key lengths and value lengths consist of the + actual lengths of individual keys/values encoded in ZeroCompressedInteger + format.
+ + @see CompressionCodec]]> +start. The starting
+ position is measured in bytes and the return value is in
+ terms of byte position in the buffer. The backing buffer is
+ not converted to a string for this operation.
+ @return byte position of the first occurrence of the search
+ string in the UTF-8 buffer or -1 if not found]]>
+ new byte[0]).]]>
+ Also includes utilities for + serializing/deserialing a string, coding/decoding a string, checking if a + byte array contains valid UTF8 code, calculating the length of an encoded + string.]]> +
DataOuput to serialize this object into.
+ @throws IOException]]>
+ For efficiency, implementations should attempt to re-use storage in the + existing object where possible.
+ + @param inDataInput to deseriablize this object from.
+ @throws IOException]]>
+ key or value type in the Hadoop Map-Reduce
+ framework implements this interface.
+
+ Implementations typically implement a static read(DataInput)
+ method which constructs a new instance, calls {@link #readFields(DataInput)}
+ and returns the instance.
Example:
+
+ public class MyWritable implements Writable {
+ // Some data
+ private int counter;
+ private long timestamp;
+
+ // Default constructor to allow (de)serialization
+ MyWritable() { }
+
+ public void write(DataOutput out) throws IOException {
+ out.writeInt(counter);
+ out.writeLong(timestamp);
+ }
+
+ public void readFields(DataInput in) throws IOException {
+ counter = in.readInt();
+ timestamp = in.readLong();
+ }
+
+ public static MyWritable read(DataInput in) throws IOException {
+ MyWritable w = new MyWritable();
+ w.readFields(in);
+ return w;
+ }
+ }
+ WritableComparables can be compared to each other, typically
+ via Comparators. Any type which is to be used as a
+ key in the Hadoop Map-Reduce framework should implement this
+ interface.
+
+ Note that hashCode() is frequently used in Hadoop to partition
+ keys. It's important that your implementation of hashCode() returns the same
+ result across different instances of the JVM. Note also that the default
+ hashCode() implementation in Object does not
+ satisfy this property.
Example:
+]]> ++ public class MyWritableComparable implements WritableComparable{ + // Some data + private int counter; + private long timestamp; + + public void write(DataOutput out) throws IOException { + out.writeInt(counter); + out.writeLong(timestamp); + } + + public void readFields(DataInput in) throws IOException { + counter = in.readInt(); + timestamp = in.readLong(); + } + + public int compareTo(MyWritableComparable o) { + int thisValue = this.value; + int thatValue = o.value; + return (thisValue < thatValue ? -1 : (thisValue==thatValue ? 0 : 1)); + } + + public int hashCode() { + final int prime = 31; + int result = 1; + result = prime * result + counter; + result = prime * result + (int) (timestamp ^ (timestamp >>> 32)); + return result + } + } +
One may optimize compare-intensive operations by overriding + {@link #compare(byte[],int,int,byte[],int,int)}. Static utility methods are + provided to assist in optimized implementations of this method.]]> +
Compressor
+ @param conf the Configuration object which contains confs for creating or reinit the compressor
+ @return Compressor for the given
+ CompressionCodec from the pool or a new one]]>
+ Decompressor
+ @return Decompressor for the given
+ CompressionCodec the pool or a new one]]>
+ b[] remain unmodified until
+ the caller is explicitly notified--via {@link #needsInput()}--that the
+ buffer may be safely modified. With this requirement, an extra
+ buffer-copy can be avoided.)
+
+ @param b Input data
+ @param off Start offset
+ @param len Length]]>
+ true if the input data buffer is empty and
+ {@link #setInput(byte[], int, int)} should be called in
+ order to provide more input.]]>
+ true if a preset dictionary is needed for decompression]]>
+ true and {@link #getRemaining()}
+ returns a positive value. finished() will be reset with the
+ {@link #reset()} method.
+ @return true if the end of the decompressed
+ data output stream has been reached.]]>
+ true and getRemaining() returns
+ a zero value, indicates that the end of data stream has been reached and
+ is not a concatenated data stream.
+ @return The number of bytes remaining in the compressed data buffer.]]>
+ false when reset() is called.]]>
+ + The behavior of TFile can be customized by the following variables through + Configuration: +
+ Suggestions on performance optimization. +
+To add a new serialization framework write an implementation of +{@link org.apache.hadoop.io.serializer.Serialization} and add its name to the +"io.serializations" property. +
]]> ++Use {@link org.apache.hadoop.io.serializer.avro.AvroSpecificSerialization} for +serialization of classes generated by Avro's 'specific' compiler. +
+ ++Use {@link org.apache.hadoop.io.serializer.avro.AvroReflectSerialization} for +other classes. +{@link org.apache.hadoop.io.serializer.avro.AvroReflectSerialization} work for +any class which is either in the package list configured via +{@link org.apache.hadoop.io.serializer.avro.AvroReflectSerialization#AVRO_REFLECT_PACKAGES} +or implement {@link org.apache.hadoop.io.serializer.avro.AvroReflectSerializable} +interface. +
]]> +basepath property. All metrics will be logged to a file in the
+ current interval's directory in a file named <hostname>.log, where
+ <hostname> is the name of the host on which the metrics logging
+ process is running. The base path is set by the
+ <prefix>.sink.<instance>.basepath property. The
+ time zone used to create the current interval's directory name is GMT. If
+ the basepath property isn't specified, it will default to
+ "/tmp", which is the temp directory on whatever default file
+ system is configured for the cluster.
+
+ The <prefix>.sink.<instance>.ignore-error
+ property controls whether an exception is thrown when an error is encountered
+ writing a log file. The default value is true. When set to
+ false, file errors are quietly swallowed.
The roll-interval property sets the amount of time before
+ rolling the directory. The default value is 1 hour. The roll interval may
+ not be less than 1 minute. The property's value should be given as
+ number unit, where number is an integer value, and
+ unit is a valid unit. Valid units are minute, hour,
+ and day. The units are case insensitive and may be abbreviated or
+ plural. If no units are specified, hours are assumed. For example,
+ "2", "2h", "2 hour", and
+ "2 hours" are all valid ways to specify two hours.
The roll-offset-interval-millis property sets the upper
+ bound on a random time interval (in milliseconds) that is used to delay
+ before the initial roll. All subsequent rolls will happen an integer
+ number of roll intervals after the initial roll, hence retaining the original
+ offset. The purpose of this property is to insert some variance in the roll
+ times so that large clusters using this sink on every node don't cause a
+ performance impact on HDFS by rolling simultaneously. The default value is
+ 30000 (30s). When writing to HDFS, as a rule of thumb, the roll offset in
+ millis should be no less than the number of sink instances times 5.
+
+
The primary use of this class is for logging to HDFS. As it uses + {@link org.apache.hadoop.fs.FileSystem} to access the target file system, + however, it can be used to write to the local file system, Amazon S3, or any + other supported file system. The base path for the sink will determine the + file system used. An unqualified path will write to the default file system + set by the configuration.
+ +Not all file systems support the ability to append to files. In file
+ systems without the ability to append to files, only one writer can write to
+ a file at a time. To allow for concurrent writes from multiple daemons on a
+ single host, the source property is used to set unique headers
+ for the log files. The property should be set to the name of
+ the source daemon, e.g. namenode. The value of the
+ source property should typically be the same as the property's
+ prefix. If this property is not set, the source is taken to be
+ unknown.
Instead of appending to an existing file, by default the sink + will create a new file with a suffix of ".<n>&quet;, where + n is the next lowest integer that isn't already used in a file name, + similar to the Hadoop daemon logs. NOTE: the file with the highest + sequence number is the newest file, unlike the Hadoop daemon logs.
+ +For file systems that allow append, the sink supports appending to the
+ existing file instead. If the allow-append property is set to
+ true, the sink will instead append to the existing file on file systems that
+ support appends. By default, the allow-append property is
+ false.
Note that when writing to HDFS with allow-append set to true,
+ there is a minimum acceptable number of data nodes. If the number of data
+ nodes drops below that minimum, the append will succeed, but reading the
+ data will fail with an IOException in the DataStreamer class. The minimum
+ number of data nodes required for a successful append is generally 2 or
+ 3.
Note also that when writing to HDFS, the file size information is not + updated until the file is closed (at the end of the interval) even though + the data is being written successfully. This is a known HDFS limitation that + exists because of the performance cost of updating the metadata. See + HDFS-5478.
+ +When using this sink in a secure (Kerberos) environment, two additional
+ properties must be set: keytab-key and
+ principal-key. keytab-key should contain the key by
+ which the keytab file can be found in the configuration, for example,
+ yarn.nodemanager.keytab. principal-key should
+ contain the key by which the principal can be found in the configuration,
+ for example, yarn.nodemanager.principal.]]>
+
+ *.sink.statsd.class=org.apache.hadoop.metrics2.sink.StatsDSink + [prefix].sink.statsd.server.host= + [prefix].sink.statsd.server.port= + [prefix].sink.statsd.skip.hostname=true|false (optional) + [prefix].sink.statsd.service.name=NameNode (name you want for service) +]]> +
Configured
+ base class, and should not be changed to do so, as it causes problems
+ for subclasses. The constructor of the Configured calls
+ the {@link #setConf(Configuration)} method, which will call into the
+ subclasses before they have been fully constructed.]]>
+ RawScriptBasedMapping that performs
+ the work: reading the configuration parameters, executing any defined
+ script, handling errors and such like. The outer
+ class extends {@link CachedDNSToSwitchMapping} to cache the delegated
+ queries.
+
+ This DNS mapper's {@link #isSingleSwitch()} predicate returns
+ true if and only if a script is defined.]]>
+ + This class uses the configuration parameter {@code + net.topology.table.file.name} to locate the mapping file. +
++ Calls to {@link #resolve(List)} will look up the address as defined in the + mapping file. If no entry corresponding to the address is found, the value + {@code /default-rack} is returned. +
]]> +null the default one will be used.]]>
+ null the default one will be used.
+ @param connConfigurator a connection configurator.]]>
+ TRUE if the token is transmitted in the
+ URL query string, FALSE if the delegation token is transmitted
+ using the {@link DelegationTokenAuthenticator#DELEGATION_TOKEN_HEADER} HTTP
+ header.]]>
+ FALSE if the delegation token is transmitted using the
+ {@link DelegationTokenAuthenticator#DELEGATION_TOKEN_HEADER} HTTP header.]]>
+ doAs parameter is not NULL,
+ the request will be done on behalf of the specified doAs user.
+
+ @param url the URL to connect to. Only HTTP/S URLs are supported.
+ @param token the authentication token being used for the user.
+ @param doAs user to do the the request on behalf of, if NULL the request is
+ as self.
+ @return an authenticated {@link HttpURLConnection}.
+ @throws IOException if an IO error occurred.
+ @throws AuthenticationException if an authentication exception occurred.]]>
+ AuthenticatedURL instances are not thread-safe.]]>
+ + This permits implementations to change the configuration before + the init operation. As the ServiceLauncher only creates + an instance of the base {@link Configuration} class, it is + recommended to instantiate any subclass (such as YarnConfiguration) + that injects new resources. +
+ @param config the initial configuration build up by the + service launcher. + @param args list of arguments passed to the command line + after any launcher-specific commands have been stripped. + @return the configuration to init the service with. + Recommended: pass down the config parameter with any changes + @throws Exception any problem]]> +
+ If an exception is raised, the policy is: +
+ Approximate HTTP equivalent: {@code 505: Version Not Supported}]]> +
+ Many of the exit codes are designed to resemble HTTP error codes, + squashed into a single byte. e.g 44 , "not found" is the equivalent + of 404. The various 2XX HTTP error codes aren't followed; + the Unix standard of "0" for success is used. +
+ 0-10: general command issues + 30-39: equivalent to the 3XX responses, where those responses are + considered errors by the application. + 40-49: client-side/CLI/config problems + 50-59: service-side problems. + 60+ : application specific error codes +]]> +
+ If the last argument is a throwable, it becomes the cause of the exception. + It will also be used as a parameter for the format. + @param exitCode exit code + @param format format for message to use in exception + @param args list of arguments]]> +
Progressable
+ to explicitly report progress to the Hadoop framework. This is especially
+ important for operations which take significant amount of time since,
+ in-lieu of the reported progress, the framework has to assume that an error
+ has occurred and time-out the operation.]]>
+ Class of the given object.]]>
+ ".sh" otherwise.
+
+ @param parent File parent directory
+ @param basename String script file basename
+ @return File referencing the script in the directory]]>
+ ".sh" otherwise.
+
+ @param basename String script file basename
+ @return String script file name]]>
+ Shell
+ processes and destroys them one by one. This method is thread safe]]>
+ + @deprecated use one of the exception-raising getter methods, + specifically {@link #getWinUtilsPath()} or {@link #getWinUtilsFile()}]]> +
du or
+ df. It also offers facilities to gate commands by
+ time-intervals.]]>
+ ShutdownHookManager singleton.]]>
+ TimeUnit]]>
+ Tool, is the standard for any Map-Reduce tool/application.
+ The tool/application should delegate the handling of
+
+ standard command-line options to {@link ToolRunner#run(Tool, String[])}
+ and only handle its custom arguments.
+
+ Here is how a typical Tool is implemented:
+ public class MyApp extends Configured implements Tool {
+
+ public int run(String[] args) throws Exception {
+ // Configuration processed by ToolRunner
+ Configuration conf = getConf();
+
+ // Create a JobConf using the processed conf
+ JobConf job = new JobConf(conf, MyApp.class);
+
+ // Process custom command-line options
+ Path in = new Path(args[1]);
+ Path out = new Path(args[2]);
+
+ // Specify various job-specific parameters
+ job.setJobName("my-app");
+ job.setInputPath(in);
+ job.setOutputPath(out);
+ job.setMapperClass(MyMapper.class);
+ job.setReducerClass(MyReducer.class);
+
+ // Submit the job, then poll for progress until the job is complete
+ RunningJob runningJob = JobClient.runJob(job);
+ if (runningJob.isSuccessful()) {
+ return 0;
+ } else {
+ return 1;
+ }
+ }
+
+ public static void main(String[] args) throws Exception {
+ // Let ToolRunner handle generic command-line options
+ int res = ToolRunner.run(new Configuration(), new MyApp(), args);
+
+ System.exit(res);
+ }
+ }
+ Configuration, or builds one if null.
+
+ Sets the Tool's configuration with the possibly modified
+ version of the conf.
+
+ @param conf Configuration for the Tool.
+ @param tool Tool to run.
+ @param args command-line arguments to the tool.
+ @return exit code of the {@link Tool#run(String[])} method.]]>
+ Configuration.
+
+ Equivalent to run(tool.getConf(), tool, args).
+
+ @param tool Tool to run.
+ @param args command-line arguments to the tool.
+ @return exit code of the {@link Tool#run(String[])} method.]]>
+ ToolRunner can be used to run classes implementing
+ Tool interface. It works in conjunction with
+ {@link GenericOptionsParser} to parse the
+
+ generic hadoop command line arguments and modifies the
+ Configuration of the Tool. The
+ application-specific options are passed along without being modified.
+
+
+ @see Tool
+ @see GenericOptionsParser]]>
+ + The Bloom filter is a data structure that was introduced in 1970 and that has been adopted by + the networking research community in the past decade thanks to the bandwidth efficiencies that it + offers for the transmission of set membership information between networked hosts. A sender encodes + the information into a bit vector, the Bloom filter, that is more compact than a conventional + representation. Computation and space costs for construction are linear in the number of elements. + The receiver uses the filter to test whether various elements are members of the set. Though the + filter will occasionally return a false positive, it will never return a false negative. When creating + the filter, the sender can choose its desired point in a trade-off between the false positive rate and the size. + +
+ Originally created by + European Commission One-Lab Project 034819. + + @see Filter The general behavior of a filter + + @see Space/Time Trade-Offs in Hash Coding with Allowable Errors]]> +
+ Invariant: nothing happens if the specified key does not belong to this counter Bloom filter. + @param key The key to remove.]]> +
NOTE: due to the bucket size of this filter, inserting the same
+ key more than 15 times will cause an overflow at all filter positions
+ associated with this key, and it will significantly increase the error
+ rate for this and other keys. For this reason the filter can only be
+ used to store small count values 0 <= N << 15.
+ @param key key to be tested
+ @return 0 if the key is not present. Otherwise, a positive value v will
+ be returned such that v == count with probability equal to the
+ error rate of this filter, and v > count otherwise.
+ Additionally, if the filter experienced an underflow as a result of
+ {@link #delete(Key)} operation, the return value may be lower than the
+ count with the probability of the false negative rate of such
+ filter.]]>
+
+ A counting Bloom filter is an improvement to standard a Bloom filter as it + allows dynamic additions and deletions of set membership information. This + is achieved through the use of a counting vector instead of a bit vector. +
+ Originally created by + European Commission One-Lab Project 034819. + + @see Filter The general behavior of a filter + + @see Summary cache: a scalable wide-area web cache sharing protocol]]> +
+ A dynamic Bloom filter (DBF) makes use of a s * m bit matrix but
+ each of the s rows is a standard Bloom filter. The creation
+ process of a DBF is iterative. At the start, the DBF is a 1 * m
+ bit matrix, i.e., it is composed of a single standard Bloom filter.
+ It assumes that nr elements are recorded in the
+ initial bit vector, where nr <= n (n is
+ the cardinality of the set A to record in the filter).
+
+ As the size of A grows during the execution of the application,
+ several keys must be inserted in the DBF. When inserting a key into the DBF,
+ one must first get an active Bloom filter in the matrix. A Bloom filter is
+ active when the number of recorded keys, nr, is
+ strictly less than the current cardinality of A, n.
+ If an active Bloom filter is found, the key is inserted and
+ nr is incremented by one. On the other hand, if there
+ is no active Bloom filter, a new one is created (i.e., a new row is added to
+ the matrix) according to the current size of A and the element
+ is added in this new Bloom filter and the nr value of
+ this new Bloom filter is set to one. A given key is said to belong to the
+ DBF if the k positions are set to one in one of the matrix rows.
+
+ Originally created by + European Commission One-Lab Project 034819. + + @see Filter The general behavior of a filter + @see BloomFilter A Bloom filter + + @see Theory and Network Applications of Dynamic Bloom Filters]]> +
+ Invariant: if the false positive is null, nothing happens.
+ @param key The false positive key to add.]]>
+
+ It allows the removal of selected false positives at the cost of introducing + random false negatives, and with the benefit of eliminating some random false + positives at the same time. + +
+ Originally created by + European Commission One-Lab Project 034819. + + @see Filter The general behavior of a filter + @see BloomFilter A Bloom filter + @see RemoveScheme The different selective clearing algorithms + + @see Retouched Bloom Filters: Allowing Networked Applications to Trade Off Selected False Positives Against False Negatives]]> +
The most important difference is that unlike GFS, Hadoop DFS files +have strictly one writer at any one time. Bytes are always appended +to the end of the writer's stream. There is no notion of "record appends" +or "mutations" that are then checked or reordered. Writers simply emit +a byte stream. That byte stream is guaranteed to be stored in the +order written.
]]> +DistributedCache is a facility provided by the Map-Reduce
+ framework to cache files (text, archives, jars etc.) needed by applications.
+
+
+ Applications specify the files, via urls (hdfs:// or http://) to be cached
+ via the {@link org.apache.hadoop.mapred.JobConf}. The
+ DistributedCache assumes that the files specified via urls are
+ already present on the {@link FileSystem} at the path specified by the url
+ and are accessible by every machine in the cluster.
The framework will copy the necessary files on to the worker node before + any tasks for the job are executed on that node. Its efficiency stems from + the fact that the files are only copied once per job and the ability to + cache archives which are un-archived on the workers.
+ +DistributedCache can be used to distribute simple, read-only
+ data/text files and/or more complex types such as archives, jars etc.
+ Archives (zip, tar and tgz/tar.gz files) are un-archived at the worker nodes.
+ Jars may be optionally added to the classpath of the tasks, a rudimentary
+ software distribution mechanism. Files have execution permissions.
+ In older version of Hadoop Map/Reduce users could optionally ask for symlinks
+ to be created in the working directory of the child task. In the current
+ version symlinks are always created. If the URL does not have a fragment
+ the name of the file or directory will be used. If multiple files or
+ directories map to the same link name, the last one added, will be used. All
+ others will not even be downloaded.
DistributedCache tracks modification timestamps of the cache
+ files. Clearly the cache files should not be modified by the application
+ or externally while the job is executing.
Here is an illustrative example on how to use the
+ DistributedCache:
+ + It is also very common to use the DistributedCache by using + {@link org.apache.hadoop.util.GenericOptionsParser}. + + This class includes methods that should be used by users + (specifically those mentioned in the example above, as well + as {@link DistributedCache#addArchiveToClassPath(Path, Configuration)}), + as well as methods intended for use by the MapReduce framework + (e.g., {@link org.apache.hadoop.mapred.JobClient}). + + @see org.apache.hadoop.mapred.JobConf + @see org.apache.hadoop.mapred.JobClient + @see org.apache.hadoop.mapreduce.Job]]> ++ // Setting up the cache for the application + + 1. Copy the requisite files to theFileSystem: + + $ bin/hadoop fs -copyFromLocal lookup.dat /myapp/lookup.dat + $ bin/hadoop fs -copyFromLocal map.zip /myapp/map.zip + $ bin/hadoop fs -copyFromLocal mylib.jar /myapp/mylib.jar + $ bin/hadoop fs -copyFromLocal mytar.tar /myapp/mytar.tar + $ bin/hadoop fs -copyFromLocal mytgz.tgz /myapp/mytgz.tgz + $ bin/hadoop fs -copyFromLocal mytargz.tar.gz /myapp/mytargz.tar.gz + + 2. Setup the application'sJobConf: + + JobConf job = new JobConf(); + DistributedCache.addCacheFile(new URI("/myapp/lookup.dat#lookup.dat"), + job); + DistributedCache.addCacheArchive(new URI("/myapp/map.zip"), job); + DistributedCache.addFileToClassPath(new Path("/myapp/mylib.jar"), job); + DistributedCache.addCacheArchive(new URI("/myapp/mytar.tar"), job); + DistributedCache.addCacheArchive(new URI("/myapp/mytgz.tgz"), job); + DistributedCache.addCacheArchive(new URI("/myapp/mytargz.tar.gz"), job); + + 3. Use the cached files in the {@link org.apache.hadoop.mapred.Mapper} + or {@link org.apache.hadoop.mapred.Reducer}: + + public static class MapClass extends MapReduceBase + implements Mapper<K, V, K, V> { + + private Path[] localArchives; + private Path[] localFiles; + + public void configure(JobConf job) { + // Get the cached archives/files + File f = new File("./map.zip/some/file/in/zip.txt"); + } + + public void map(K key, V value, + OutputCollector<K, V> output, Reporter reporter) + throws IOException { + // Use data from the cached archives/files here + // ... + // ... + output.collect(k, v); + } + } + +
JobTracker.]]>
+ ClusterStatus provides clients with information such as:
+ JobTracker.
+ Clients can query for the latest ClusterStatus, via
+ {@link JobClient#getClusterStatus()}.
Counters represent global counters, defined either by the
+ Map-Reduce framework or applications. Each Counter can be of
+ any {@link Enum} type.
+
+ Counters are bunched into {@link Group}s, each comprising of
+ counters from a particular Enum class.]]>
+
Grouphandles localization of the class name and the
+ counter names.
FileInputFormat implementations can override this and return
+ false to ensure that individual input files are never split-up
+ so that {@link Mapper}s process entire files.
+
+ @param fs the file system that the file is on
+ @param filename the file name to check
+ @return is this file splitable?]]>
+ FileInputFormat is the base class for all file-based
+ InputFormats. This provides a generic implementation of
+ {@link #getSplits(JobConf, int)}.
+
+ Implementations of FileInputFormat can also override the
+ {@link #isSplitable(FileSystem, Path)} method to prevent input files
+ from being split-up in certain situations. Implementations that may
+ deal with non-splittable files must override this method, since
+ the default implementation assumes splitting is always possible.]]>
+ false otherwise]]>
+ Note: The following is valid only if the {@link OutputCommitter}
+ is {@link FileOutputCommitter}. If OutputCommitter is not
+ a FileOutputCommitter, the task's temporary output
+ directory is same as {@link #getOutputPath(JobConf)} i.e.
+ ${mapreduce.output.fileoutputformat.outputdir}$
Some applications need to create/write-to side-files, which differ from + the actual job-outputs. + +
In such cases there could be issues with 2 instances of the same TIP + (running simultaneously e.g. speculative tasks) trying to open/write-to the + same file (path) on HDFS. Hence the application-writer will have to pick + unique names per task-attempt (e.g. using the attemptid, say + attempt_200709221812_0001_m_000000_0), not just per TIP.
+ +To get around this the Map-Reduce framework helps the application-writer + out by maintaining a special + ${mapreduce.output.fileoutputformat.outputdir}/_temporary/_${taskid} + sub-directory for each task-attempt on HDFS where the output of the + task-attempt goes. On successful completion of the task-attempt the files + in the ${mapreduce.output.fileoutputformat.outputdir}/_temporary/_${taskid} (only) + are promoted to ${mapreduce.output.fileoutputformat.outputdir}. Of course, the + framework discards the sub-directory of unsuccessful task-attempts. This + is completely transparent to the application.
+ +The application-writer can take advantage of this by creating any + side-files required in ${mapreduce.task.output.dir} during execution + of his reduce-task i.e. via {@link #getWorkOutputPath(JobConf)}, and the + framework will move them out similarly - thus she doesn't have to pick + unique paths per task-attempt.
+ +Note: the value of ${mapreduce.task.output.dir} during + execution of a particular task-attempt is actually + ${mapreduce.output.fileoutputformat.outputdir}/_temporary/_{$taskid}, and this value is + set by the map-reduce framework. So, just create any side-files in the + path returned by {@link #getWorkOutputPath(JobConf)} from map/reduce + task to take advantage of this feature.
+ +The entire discussion holds true for maps of jobs with + reducer=NONE (i.e. 0 reduces) since output of the map, in that case, + goes directly to HDFS.
+ + @return the {@link Path} to the task's temporary output directory + for the map-reduce job.]]> +The given name is postfixed with the task type, 'm' for maps, 'r' for + reduces and the task partition number. For example, give a name 'test' + running on the first map o the job the generated name will be + 'test-m-00000'.
+ + @param conf the configuration for the job. + @param name the name to make unique. + @return a unique name accross all tasks of the job.]]> +This method uses the {@link #getUniqueName} method to make the file name + unique for the task.
+ + @param conf the configuration for the job. + @param name the name for the file. + @return a unique path accross all tasks of the job.]]> +Note: The split is a logical split of the inputs and the + input files are not physically split into chunks. For e.g. a split could + be <input-file-path, start, offset> tuple. + + @param job job configuration. + @param numSplits the desired number of splits, a hint. + @return an array of {@link InputSplit}s for the job.]]> +
RecordReader to respect
+ record boundaries while processing the logical split to present a
+ record-oriented view to the individual task.
+
+ @param split the {@link InputSplit}
+ @param job the job that this split belongs to
+ @return a {@link RecordReader}]]>
+ The Map-Reduce framework relies on the InputFormat of the
+ job to:
+
InputSplit for processing by
+ the {@link Mapper}.
+ The default behavior of file-based {@link InputFormat}s, typically + sub-classes of {@link FileInputFormat}, is to split the + input into logical {@link InputSplit}s based on the total size, in + bytes, of the input files. However, the {@link FileSystem} blocksize of + the input files is treated as an upper bound for input splits. A lower bound + on the split size can be set via + + mapreduce.input.fileinputformat.split.minsize.
+ +Clearly, logical splits based on input-size is insufficient for many
+ applications since record boundaries are to be respected. In such cases, the
+ application has to also implement a {@link RecordReader} on whom lies the
+ responsibilty to respect record-boundaries and present a record-oriented
+ view of the logical InputSplit to the individual task.
+
+ @see InputSplit
+ @see RecordReader
+ @see JobClient
+ @see FileInputFormat]]>
+
Strings.
+ @throws IOException]]>
+ Typically, it presents a byte-oriented view on the input and is the + responsibility of {@link RecordReader} of the job to process this and present + a record-oriented view. + + @see InputFormat + @see RecordReader]]> +
JobClient provides facilities to submit jobs, track their
+ progress, access component-tasks' reports/logs, get the Map-Reduce cluster
+ status information etc.
+
+ The job submission process involves: +
JobClient to submit
+ the job and monitor its progress.
+
+ Here is an example on how to use JobClient:
+ // Create a new JobConf
+ JobConf job = new JobConf(new Configuration(), MyJob.class);
+
+ // Specify various job-specific parameters
+ job.setJobName("myjob");
+
+ job.setInputPath(new Path("in"));
+ job.setOutputPath(new Path("out"));
+
+ job.setMapperClass(MyJob.MyMapper.class);
+ job.setReducerClass(MyJob.MyReducer.class);
+
+ // Submit the job, then poll for progress until the job is complete
+ JobClient.runJob(job);
+ At times clients would chain map-reduce jobs to accomplish complex tasks + which cannot be done via a single map-reduce job. This is fairly easy since + the output of the job, typically, goes to distributed file-system and that + can be used as the input for the next job.
+ +However, this also means that the onus on ensuring jobs are complete + (success/failure) lies squarely on the clients. In such situations the + various job-control options are: +
false otherwise.]]>
+ false otherwise.]]>
+ For key-value pairs (K1,V1) and (K2,V2), the values (V1, V2) are passed + in a single call to the reduce function if K1 and K2 compare as equal.
+ +Since {@link #setOutputKeyComparatorClass(Class)} can be used to control + how keys are sorted, this can be used in conjunction to simulate + secondary sort on values.
+ +Note: This is not a guarantee of the combiner sort being + stable in any sense. (In any case, with the order of available + map-outputs to the combiner being non-deterministic, it wouldn't make + that much sense.)
+ + @param theClass the comparator class to be used for grouping keys for the + combiner. It should implementRawComparator.
+ @see #setOutputKeyComparatorClass(Class)]]>
+ For key-value pairs (K1,V1) and (K2,V2), the values (V1, V2) are passed + in a single call to the reduce function if K1 and K2 compare as equal.
+ +Since {@link #setOutputKeyComparatorClass(Class)} can be used to control + how keys are sorted, this can be used in conjunction to simulate + secondary sort on values.
+ +Note: This is not a guarantee of the reduce sort being + stable in any sense. (In any case, with the order of available + map-outputs to the reduce being non-deterministic, it wouldn't make + that much sense.)
+ + @param theClass the comparator class to be used for grouping keys. + It should implementRawComparator.
+ @see #setOutputKeyComparatorClass(Class)
+ @see #setCombinerKeyGroupingComparator(Class)]]>
+ The combiner is an application-specified aggregation operation, which + can help cut down the amount of data transferred between the + {@link Mapper} and the {@link Reducer}, leading to better performance.
+ +The framework may invoke the combiner 0, 1, or multiple times, in both + the mapper and reducer tasks. In general, the combiner is called as the + sort/merge result is written to disk. The combiner must: +
Typically the combiner is same as the Reducer for the
+ job i.e. {@link #setReducerClass(Class)}.
true if speculative execution be used for this job,
+ false otherwise.]]>
+ false.]]>
+ true if speculative execution be
+ used for this job for map tasks,
+ false otherwise.]]>
+ false.]]>
+ true if speculative execution be used
+ for reduce tasks for this job,
+ false otherwise.]]>
+ false.]]>
+ The number of maps is usually driven by the total size of the inputs + i.e. total number of blocks of the input files.
+ +The right level of parallelism for maps seems to be around 10-100 maps + per-node, although it has been set up to 300 or so for very cpu-light map + tasks. Task setup takes awhile, so it is best if the maps take at least a + minute to execute.
+ +The default behavior of file-based {@link InputFormat}s is to split the + input into logical {@link InputSplit}s based on the total size, in + bytes, of input files. However, the {@link FileSystem} blocksize of the + input files is treated as an upper bound for input splits. A lower bound + on the split size can be set via + + mapreduce.input.fileinputformat.split.minsize.
+ +Thus, if you expect 10TB of input data and have a blocksize of 128MB, + you'll end up with 82,000 maps, unless {@link #setNumMapTasks(int)} is + used to set it even higher.
+ + @param n the number of map tasks for this job. + @see InputFormat#getSplits(JobConf, int) + @see FileInputFormat + @see FileSystem#getDefaultBlockSize() + @see FileStatus#getBlockSize()]]> +The right number of reduces seems to be 0.95 or
+ 1.75 multiplied by (
+ available memory for reduce tasks
+ (The value of this should be smaller than
+ numNodes * yarn.nodemanager.resource.memory-mb
+ since the resource of memory is shared by map tasks and other
+ applications) /
+
+ mapreduce.reduce.memory.mb).
+
With 0.95 all of the reduces can launch immediately and
+ start transfering map outputs as the maps finish. With 1.75
+ the faster nodes will finish their first round of reduces and launch a
+ second wave of reduces doing a much better job of load balancing.
Increasing the number of reduces increases the framework overhead, but + increases load balancing and lowers the cost of failures.
+ +The scaling factors above are slightly less than whole numbers to + reserve a few reduce slots in the framework for speculative-tasks, failures + etc.
+ + Reducer NONE + +It is legal to set the number of reduce-tasks to zero.
In this case the output of the map-tasks directly go to distributed + file-system, to the path set by + {@link FileOutputFormat#setOutputPath(JobConf, Path)}. Also, the + framework doesn't sort the map-outputs before writing it out to HDFS.
+ + @param n the number of reduce tasks for this job.]]> +zero, i.e. any failed map-task results in
+ the job being declared as {@link JobStatus#FAILED}.
+
+ @return the maximum percentage of map tasks that can fail without
+ the job being aborted.]]>
+ zero, i.e. any failed reduce-task results
+ in the job being declared as {@link JobStatus#FAILED}.
+
+ @return the maximum percentage of reduce tasks that can fail without
+ the job being aborted.]]>
+ The debug command, run on the node where the map failed, is:
++ ++ $script $stdout $stderr $syslog $jobconf. +
The script file is distributed through {@link DistributedCache} + APIs. The script needs to be symlinked.
+ +Here is an example on how to submit a script +
+ job.setMapDebugScript("./myscript");
+ DistributedCache.createSymlink(job);
+ DistributedCache.addCacheFile("/debug/scripts/myscript#myscript");
+ The debug command, run on the node where the map failed, is:
++ ++ $script $stdout $stderr $syslog $jobconf. +
The script file is distributed through {@link DistributedCache} + APIs. The script file needs to be symlinked
+ +Here is an example on how to submit a script +
+ job.setReduceDebugScript("./myscript");
+ DistributedCache.createSymlink(job);
+ DistributedCache.addCacheFile("/debug/scripts/myscript#myscript");
+ This is typically used by application-writers to implement chaining of + Map-Reduce jobs in an asynchronous manner.
+ + @param uri the job end notification uri + @see JobStatus]]> +
+ ${mapreduce.cluster.local.dir}/taskTracker/$user/jobcache/$jobid/work/ .
+ This directory is exposed to the users through
+ mapreduce.job.local.dir .
+ So, the tasks can use this space
+ as scratch space and share files among them.
+ This value is available as System property also.
+
+ @return The localized job specific shared directory]]>
+ + For backward compatibility, if the job configuration sets the + key {@link #MAPRED_TASK_MAXVMEM_PROPERTY}, that value is returned. + Otherwise, this method will return the larger of the values returned by + {@link #getMemoryForMapTask()} and {@link #getMemoryForReduceTask()} + after converting them into bytes. + + @return Memory required to run a task of this job, in bytes. + @see #setMaxVirtualMemoryForTask(long) + @deprecated Use {@link #getMemoryForMapTask()} and + {@link #getMemoryForReduceTask()}]]> +
$key on
+ Linux or %key% on Windows.
+
+ Example:
+ $key on
+ Linux or %key% on Windows.
+
+ Example:
+ $key on
+ Linux or %key% on Windows.
+
+ Example:
+ JobConf is the primary interface for a user to describe a
+ map-reduce job to the Hadoop framework for execution. The framework tries to
+ faithfully execute the job as-is described by JobConf, however:
+ JobConf typically specifies the {@link Mapper}, combiner
+ (if any), {@link Partitioner}, {@link Reducer}, {@link InputFormat} and
+ {@link OutputFormat} implementations to be used etc.
+
+
Optionally JobConf is used to specify other advanced facets
+ of the job such as Comparators to be used, files to be put in
+ the {@link DistributedCache}, whether or not intermediate and/or job outputs
+ are to be compressed (and how), debugability via user-provided scripts
+ ( {@link #setMapDebugScript(String)}/{@link #setReduceDebugScript(String)}),
+ for doing post-processing on task logs, task's stdout, stderr, syslog.
+ and etc.
Here is an example on how to configure a job via JobConf:
+ // Create a new JobConf
+ JobConf job = new JobConf(new Configuration(), MyJob.class);
+
+ // Specify various job-specific parameters
+ job.setJobName("myjob");
+
+ FileInputFormat.setInputPaths(job, new Path("in"));
+ FileOutputFormat.setOutputPath(job, new Path("out"));
+
+ job.setMapperClass(MyJob.MyMapper.class);
+ job.setCombinerClass(MyJob.MyReducer.class);
+ job.setReducerClass(MyJob.MyReducer.class);
+
+ job.setInputFormat(SequenceFileInputFormat.class);
+ job.setOutputFormat(SequenceFileOutputFormat.class);
+
+ JobID.getTaskIDsPattern("200707121733", null);
+
+ which will return :
+ "job_200707121733_[0-9]*"+ @param jtIdentifier jobTracker identifier, or null + @param jobId job number, or null + @return a regex pattern matching JobIDs]]> +
job_200707121733_0003 , which represents the third job
+ running at the jobtracker started at 200707121733.
+ + Applications should never construct or parse JobID strings, but rather + use appropriate constructors or {@link #forName(String)} method. + + @see TaskID + @see TaskAttemptID]]> +
Applications can use the {@link Reporter} provided to report progress + or just indicate that they are alive. In scenarios where the application + takes significant amount of time to process individual key/value + pairs, this is crucial since the framework might assume that the task has + timed-out and kill that task. The other way of avoiding this is to set + + mapreduce.task.timeout to a high-enough value (or even zero for no + time-outs).
+ + @param key the input key. + @param value the input value. + @param output collects mapped keys and values. + @param reporter facility to report progress.]]> +The Hadoop Map-Reduce framework spawns one map task for each
+ {@link InputSplit} generated by the {@link InputFormat} for the job.
+ Mapper implementations can access the {@link JobConf} for the
+ job via the {@link JobConfigurable#configure(JobConf)} and initialize
+ themselves. Similarly they can use the {@link Closeable#close()} method for
+ de-initialization.
The framework then calls
+ {@link #map(Object, Object, OutputCollector, Reporter)}
+ for each key/value pair in the InputSplit for that task.
All intermediate values associated with a given output key are
+ subsequently grouped by the framework, and passed to a {@link Reducer} to
+ determine the final output. Users can control the grouping by specifying
+ a Comparator via
+ {@link JobConf#setOutputKeyComparatorClass(Class)}.
The grouped Mapper outputs are partitioned per
+ Reducer. Users can control which keys (and hence records) go to
+ which Reducer by implementing a custom {@link Partitioner}.
+
+
Users can optionally specify a combiner, via
+ {@link JobConf#setCombinerClass(Class)}, to perform local aggregation of the
+ intermediate outputs, which helps to cut down the amount of data transferred
+ from the Mapper to the Reducer.
+
+
The intermediate, grouped outputs are always stored in
+ {@link SequenceFile}s. Applications can specify if and how the intermediate
+ outputs are to be compressed and which {@link CompressionCodec}s are to be
+ used via the JobConf.
If the job has
+ zero
+ reduces then the output of the Mapper is directly written
+ to the {@link FileSystem} without grouping by keys.
Example:
+
+ public class MyMapper<K extends WritableComparable, V extends Writable>
+ extends MapReduceBase implements Mapper<K, V, K, V> {
+
+ static enum MyCounters { NUM_RECORDS }
+
+ private String mapTaskId;
+ private String inputFile;
+ private int noRecords = 0;
+
+ public void configure(JobConf job) {
+ mapTaskId = job.get(JobContext.TASK_ATTEMPT_ID);
+ inputFile = job.get(JobContext.MAP_INPUT_FILE);
+ }
+
+ public void map(K key, V val,
+ OutputCollector<K, V> output, Reporter reporter)
+ throws IOException {
+ // Process the <key, value> pair (assume this takes a while)
+ // ...
+ // ...
+
+ // Let the framework know that we are alive, and kicking!
+ // reporter.progress();
+
+ // Process some more
+ // ...
+ // ...
+
+ // Increment the no. of <key, value> pairs processed
+ ++noRecords;
+
+ // Increment counters
+ reporter.incrCounter(NUM_RECORDS, 1);
+
+ // Every 100 records update application-level status
+ if ((noRecords%100) == 0) {
+ reporter.setStatus(mapTaskId + " processed " + noRecords +
+ " from input-file: " + inputFile);
+ }
+
+ // Output the result
+ output.collect(key, val);
+ }
+ }
+ Applications may write a custom {@link MapRunnable} to exert greater
+ control on map processing e.g. multi-threaded Mappers etc.
Mapping of input records to output records is complete when this method + returns.
+ + @param input the {@link RecordReader} to read the input records. + @param output the {@link OutputCollector} to collect the outputrecords. + @param reporter {@link Reporter} to report progress, status-updates etc. + @throws IOException]]> +MapRunnable can exert greater
+ control on map processing e.g. multi-threaded, asynchronous mappers etc.
+
+ @see Mapper]]>
+ RecordReader's for MultiFileSplit's.
+ @see MultiFileSplit]]>
+ OutputCollector is the generalization of the facility
+ provided by the Map-Reduce framework to collect data output by either the
+ Mapper or the Reducer i.e. intermediate outputs
+ or the output of the job.
false otherwise
+ @throws IOException
+ @see #recoverTask(TaskAttemptContext)]]>
+ false otherwise
+ @throws IOException]]>
+ The Map-Reduce framework relies on the OutputCommitter of
+ the job to:
+
The Map-Reduce framework relies on the OutputFormat of the
+ job to:
+
key.]]>
+ Partitioner controls the partitioning of the keys of the
+ intermediate map-outputs. The key (or a subset of the key) is used to derive
+ the partition, typically by a hash function. The total number of partitions
+ is the same as the number of reduce tasks for the job. Hence this controls
+ which of the m reduce tasks the intermediate key (and hence the
+ record) is sent for reduction.
+
+ Note: A Partitioner is created only when there are multiple
+ reducers.
1.0.
+ @throws IOException]]>
+ RecordReader, typically, converts the byte-oriented view of
+ the input, provided by the InputSplit, and presents a
+ record-oriented view for the {@link Mapper} and {@link Reducer} tasks for
+ processing. It thus assumes the responsibility of processing record
+ boundaries and presenting the tasks with keys and values.
RecordWriter implementations write the job outputs to the
+ {@link FileSystem}.
+
+ @see OutputFormat]]>
+
The framework calls this method for each
+ <key, (list of values)> pair in the grouped inputs.
+ Output values must be of the same type as input values. Input keys must
+ not be altered. The framework will reuse the key and value objects
+ that are passed into the reduce, therefore the application should clone
+ the objects they want to keep a copy of. In many cases, all values are
+ combined into zero or one value.
+
Output pairs are collected with calls to + {@link OutputCollector#collect(Object,Object)}.
+ +Applications can use the {@link Reporter} provided to report progress + or just indicate that they are alive. In scenarios where the application + takes a significant amount of time to process individual key/value + pairs, this is crucial since the framework might assume that the task has + timed-out and kill that task. The other way of avoiding this is to set + + mapreduce.task.timeout to a high-enough value (or even zero for no + time-outs).
+ + @param key the key. + @param values the list of values to reduce. + @param output to collect keys and combined values. + @param reporter facility to report progress.]]> +Reducers for the job is set by the user via
+ {@link JobConf#setNumReduceTasks(int)}. Reducer implementations
+ can access the {@link JobConf} for the job via the
+ {@link JobConfigurable#configure(JobConf)} method and initialize themselves.
+ Similarly they can use the {@link Closeable#close()} method for
+ de-initialization.
+
+ Reducer has 3 primary phases:
Reducer is input the grouped output of a {@link Mapper}.
+ In the phase the framework, for each Reducer, fetches the
+ relevant partition of the output of all the Mappers, via HTTP.
+
The framework groups Reducer inputs by keys
+ (since different Mappers may have output the same key) in this
+ stage.
The shuffle and sort phases occur simultaneously i.e. while outputs are + being fetched they are merged.
+ + SecondarySort + +If equivalence rules for keys while grouping the intermediates are
+ different from those for grouping keys before reduction, then one may
+ specify a Comparator via
+ {@link JobConf#setOutputValueGroupingComparator(Class)}.Since
+ {@link JobConf#setOutputKeyComparatorClass(Class)} can be used to
+ control how intermediate keys are grouped, these can be used in conjunction
+ to simulate secondary sort on values.
In this phase the
+ {@link #reduce(Object, Iterator, OutputCollector, Reporter)}
+ method is called for each <key, (list of values)> pair in
+ the grouped inputs.
The output of the reduce task is typically written to the + {@link FileSystem} via + {@link OutputCollector#collect(Object, Object)}.
+The output of the Reducer is not re-sorted.
Example:
+
+ public class MyReducer<K extends WritableComparable, V extends Writable>
+ extends MapReduceBase implements Reducer<K, V, K, V> {
+
+ static enum MyCounters { NUM_RECORDS }
+
+ private String reduceTaskId;
+ private int noKeys = 0;
+
+ public void configure(JobConf job) {
+ reduceTaskId = job.get(JobContext.TASK_ATTEMPT_ID);
+ }
+
+ public void reduce(K key, Iterator<V> values,
+ OutputCollector<K, V> output,
+ Reporter reporter)
+ throws IOException {
+
+ // Process
+ int noValues = 0;
+ while (values.hasNext()) {
+ V value = values.next();
+
+ // Increment the no. of values for this key
+ ++noValues;
+
+ // Process the <key, value> pair (assume this takes a while)
+ // ...
+ // ...
+
+ // Let the framework know that we are alive, and kicking!
+ if ((noValues%10) == 0) {
+ reporter.progress();
+ }
+
+ // Process some more
+ // ...
+ // ...
+
+ // Output the <key, value>
+ output.collect(key, value);
+ }
+
+ // Increment the no. of <key, list of values> pairs processed
+ ++noKeys;
+
+ // Increment counters
+ reporter.incrCounter(NUM_RECORDS, 1);
+
+ // Every 100 keys update application-level status
+ if ((noKeys%100) == 0) {
+ reporter.setStatus(reduceTaskId + " processed " + noKeys);
+ }
+ }
+ }
+ Reporter
+ provided to report progress or just indicate that they are alive. In
+ scenarios where the application takes significant amount of time to
+ process individual key/value pairs, this is crucial since the framework
+ might assume that the task has timed-out and kill that task.
+
+ Applications can also update {@link Counters} via the provided
+ Reporter .
false.
+ @throws IOException]]>
+ false.
+ @throws IOException]]>
+ false.
+ @throws IOException]]>
+ Clients can get hold of RunningJob via the {@link JobClient}
+ and then query the running-job for details such as name, configuration,
+ progress etc.
false otherwise.]]>
+ false otherwise.]]>
+ This feature can be used when map/reduce tasks crashes deterministically on + certain input. This happens due to bugs in the map/reduce function. The usual + course would be to fix these bugs. But sometimes this is not possible; + perhaps the bug is in third party libraries for which the source code is + not available. Due to this, the task never reaches to completion even with + multiple attempts and complete data for that task is lost.
+ +With this feature, only a small portion of data is lost surrounding + the bad record, which may be acceptable for some user applications. + see {@link SkipBadRecords#setMapperMaxSkipRecords(Configuration, long)}
+ +The skipping mode gets kicked off after certain no of failures + see {@link SkipBadRecords#setAttemptsToStartSkipping(Configuration, int)}
+ +In the skipping mode, the map/reduce task maintains the record range which + is getting processed at all times. Before giving the input to the + map/reduce function, it sends this record range to the Task tracker. + If task crashes, the Task tracker knows which one was the last reported + range. On further attempts that range get skipped.
]]> ++ TaskAttemptID.getTaskAttemptIDsPattern(null, null, true, 1, null); ++ which will return : +
"attempt_[^_]*_[0-9]*_m_000001_[0-9]*"+ @param jtIdentifier jobTracker identifier, or null + @param jobId job number, or null + @param isMap whether the tip is a map, or null + @param taskId taskId number, or null + @param attemptId the task attempt number, or null + @return a regex pattern matching TaskAttemptIDs]]> +
+ TaskAttemptID.getTaskAttemptIDsPattern(null, null, TaskType.MAP, 1, null); ++ which will return : +
"attempt_[^_]*_[0-9]*_m_000001_[0-9]*"+ @param jtIdentifier jobTracker identifier, or null + @param jobId job number, or null + @param type the {@link TaskType} + @param taskId taskId number, or null + @param attemptId the task attempt number, or null + @return a regex pattern matching TaskAttemptIDs]]> +
attempt_200707121733_0003_m_000005_0 , which represents the
+ zeroth task attempt for the fifth map task in the third job
+ running at the jobtracker started at 200707121733.
+ + Applications should never construct or parse TaskAttemptID strings + , but rather use appropriate constructors or {@link #forName(String)} + method. + + @see JobID + @see TaskID]]> +
+ TaskID.getTaskIDsPattern(null, null, true, 1); ++ which will return : +
"task_[^_]*_[0-9]*_m_000001*"+ @param jtIdentifier jobTracker identifier, or null + @param jobId job number, or null + @param isMap whether the tip is a map, or null + @param taskId taskId number, or null + @return a regex pattern matching TaskIDs + @deprecated Use {@link TaskID#getTaskIDsPattern(String, Integer, TaskType, + Integer)}]]> +
+ TaskID.getTaskIDsPattern(null, null, true, 1); ++ which will return : +
"task_[^_]*_[0-9]*_m_000001*"+ @param jtIdentifier jobTracker identifier, or null + @param jobId job number, or null + @param type the {@link TaskType}, or null + @param taskId taskId number, or null + @return a regex pattern matching TaskIDs]]> +
task_200707121733_0003_m_000005 , which represents the
+ fifth map task in the third job running at the jobtracker
+ started at 200707121733.
+ + Applications should never construct or parse TaskID strings + , but rather use appropriate constructors or {@link #forName(String)} + method. + + @see JobID + @see TaskAttemptID]]> +
) }]]> +
+ For the added Mapper the configuration given for it,
+ mapperConf, have precedence over the job's JobConf. This
+ precedence is in effect when the task is running.
+
+ IMPORTANT: There is no need to specify the output key/value classes for the + ChainMapper, this is done by the addMapper for the last mapper in the chain +
+
+ @param job job's JobConf to add the Mapper class.
+ @param klass the Mapper class to add.
+ @param inputKeyClass mapper input key class.
+ @param inputValueClass mapper input value class.
+ @param outputKeyClass mapper output key class.
+ @param outputValueClass mapper output value class.
+ @param byValue indicates if key/values should be passed by value
+ to the next Mapper in the chain, if any.
+ @param mapperConf a JobConf with the configuration for the Mapper
+ class. It is recommended to use a JobConf without default values using the
+ JobConf(boolean loadDefaults) constructor with FALSE.]]>
+
super.configure(...) should be
+ invoked at the beginning of the overwriter method.]]>
+ super.close() should be
+ invoked at the end of the overwriter method.]]>
+ + The key functionality of this feature is that the Mappers in the chain do not + need to be aware that they are executed in a chain. This enables having + reusable specialized Mappers that can be combined to perform composite + operations within a single task. +
+ Special care has to be taken when creating chains that the key/values output + by a Mapper are valid for the following Mapper in the chain. It is assumed + all Mappers and the Reduce in the chain use maching output and input key and + value classes as no conversion is done by the chaining code. +
+ Using the ChainMapper and the ChainReducer classes is possible to compose
+ Map/Reduce jobs that look like [MAP+ / REDUCE MAP*]. And
+ immediate benefit of this pattern is a dramatic reduction in disk IO.
+
+ IMPORTANT: There is no need to specify the output key/value classes for the + ChainMapper, this is done by the addMapper for the last mapper in the chain. +
+ ChainMapper usage pattern: +
+
+ ...
+ conf.setJobName("chain");
+ conf.setInputFormat(TextInputFormat.class);
+ conf.setOutputFormat(TextOutputFormat.class);
+
+ JobConf mapAConf = new JobConf(false);
+ ...
+ ChainMapper.addMapper(conf, AMap.class, LongWritable.class, Text.class,
+ Text.class, Text.class, true, mapAConf);
+
+ JobConf mapBConf = new JobConf(false);
+ ...
+ ChainMapper.addMapper(conf, BMap.class, Text.class, Text.class,
+ LongWritable.class, Text.class, false, mapBConf);
+
+ JobConf reduceConf = new JobConf(false);
+ ...
+ ChainReducer.setReducer(conf, XReduce.class, LongWritable.class, Text.class,
+ Text.class, Text.class, true, reduceConf);
+
+ ChainReducer.addMapper(conf, CMap.class, Text.class, Text.class,
+ LongWritable.class, Text.class, false, null);
+
+ ChainReducer.addMapper(conf, DMap.class, LongWritable.class, Text.class,
+ LongWritable.class, LongWritable.class, true, null);
+
+ FileInputFormat.setInputPaths(conf, inDir);
+ FileOutputFormat.setOutputPath(conf, outDir);
+ ...
+
+ JobClient jc = new JobClient(conf);
+ RunningJob job = jc.submitJob(conf);
+ ...
+ ]]>
+
+ For the added Reducer the configuration given for it,
+ reducerConf, have precedence over the job's JobConf. This
+ precedence is in effect when the task is running.
+
+ IMPORTANT: There is no need to specify the output key/value classes for the
+ ChainReducer, this is done by the setReducer or the addMapper for the last
+ element in the chain.
+
+ @param job job's JobConf to add the Reducer class.
+ @param klass the Reducer class to add.
+ @param inputKeyClass reducer input key class.
+ @param inputValueClass reducer input value class.
+ @param outputKeyClass reducer output key class.
+ @param outputValueClass reducer output value class.
+ @param byValue indicates if key/values should be passed by value
+ to the next Mapper in the chain, if any.
+ @param reducerConf a JobConf with the configuration for the Reducer
+ class. It is recommended to use a JobConf without default values using the
+ JobConf(boolean loadDefaults) constructor with FALSE.]]>
+
+ For the added Mapper the configuration given for it,
+ mapperConf, have precedence over the job's JobConf. This
+ precedence is in effect when the task is running.
+
+ IMPORTANT: There is no need to specify the output key/value classes for the
+ ChainMapper, this is done by the addMapper for the last mapper in the chain
+ .
+
+ @param job chain job's JobConf to add the Mapper class.
+ @param klass the Mapper class to add.
+ @param inputKeyClass mapper input key class.
+ @param inputValueClass mapper input value class.
+ @param outputKeyClass mapper output key class.
+ @param outputValueClass mapper output value class.
+ @param byValue indicates if key/values should be passed by value
+ to the next Mapper in the chain, if any.
+ @param mapperConf a JobConf with the configuration for the Mapper
+ class. It is recommended to use a JobConf without default values using the
+ JobConf(boolean loadDefaults) constructor with FALSE.]]>
+
super.configure(...) should be
+ invoked at the beginning of the overwriter method.]]>
+ map(...) methods of the Mappers in the chain.]]>
+ super.close() should be
+ invoked at the end of the overwriter method.]]>
+ + The key functionality of this feature is that the Mappers in the chain do not + need to be aware that they are executed after the Reducer or in a chain. + This enables having reusable specialized Mappers that can be combined to + perform composite operations within a single task. +
+ Special care has to be taken when creating chains that the key/values output + by a Mapper are valid for the following Mapper in the chain. It is assumed + all Mappers and the Reduce in the chain use maching output and input key and + value classes as no conversion is done by the chaining code. +
+ Using the ChainMapper and the ChainReducer classes is possible to compose
+ Map/Reduce jobs that look like [MAP+ / REDUCE MAP*]. And
+ immediate benefit of this pattern is a dramatic reduction in disk IO.
+
+ IMPORTANT: There is no need to specify the output key/value classes for the + ChainReducer, this is done by the setReducer or the addMapper for the last + element in the chain. +
+ ChainReducer usage pattern: +
+
+ ...
+ conf.setJobName("chain");
+ conf.setInputFormat(TextInputFormat.class);
+ conf.setOutputFormat(TextOutputFormat.class);
+
+ JobConf mapAConf = new JobConf(false);
+ ...
+ ChainMapper.addMapper(conf, AMap.class, LongWritable.class, Text.class,
+ Text.class, Text.class, true, mapAConf);
+
+ JobConf mapBConf = new JobConf(false);
+ ...
+ ChainMapper.addMapper(conf, BMap.class, Text.class, Text.class,
+ LongWritable.class, Text.class, false, mapBConf);
+
+ JobConf reduceConf = new JobConf(false);
+ ...
+ ChainReducer.setReducer(conf, XReduce.class, LongWritable.class, Text.class,
+ Text.class, Text.class, true, reduceConf);
+
+ ChainReducer.addMapper(conf, CMap.class, Text.class, Text.class,
+ LongWritable.class, Text.class, false, null);
+
+ ChainReducer.addMapper(conf, DMap.class, LongWritable.class, Text.class,
+ LongWritable.class, LongWritable.class, true, null);
+
+ FileInputFormat.setInputPaths(conf, inDir);
+ FileOutputFormat.setOutputPath(conf, outDir);
+ ...
+
+ JobClient jc = new JobClient(conf);
+ RunningJob job = jc.submitJob(conf);
+ ...
+ ]]>
+ CombineFileSplit's.
+ @see CombineFileSplit]]>
+ SequenceFileInputFormat.
+
+ @see CombineFileInputFormat]]>
+ TextInputFormat.
+
+ @see CombineFileInputFormat]]>
+ false
+ if it is single. If the name output is not defined it returns
+ false]]>
+ + MultipleOutputs supports counters, by default the are disabled. + The counters group is the {@link MultipleOutputs} class name. +
+ The names of the counters are the same as the named outputs. For multi + named outputs the name of the counter is the concatenation of the named + output, and underscore '_' and the multiname. + + @param conf job conf to enableadd the named output. + @param enabled indicates if the counters will be enabled or not.]]> ++ MultipleOutputs supports counters, by default the are disabled. + The counters group is the {@link MultipleOutputs} class name. +
+ The names of the counters are the same as the named outputs. For multi + named outputs the name of the counter is the concatenation of the named + output, and underscore '_' and the multiname. + + + @param conf job conf to enableadd the named output. + @return TRUE if the counters are enabled, FALSE if they are disabled.]]> +super.close() at the
+ end of their close()
+
+ @throws java.io.IOException thrown if any of the MultipleOutput files
+ could not be closed properly.]]>
+ map() and reduce() methods of the
+ Mapper and Reducer implementations.
+
+ Each additional output, or named output, may be configured with its own
+ OutputFormat, with its own key class and with its own value
+ class.
+
+ A named output can be a single file or a multi file. The later is referred as + a multi named output. +
+ A multi named output is an unbound set of files all sharing the same
+ OutputFormat, key class and value class configuration.
+
+ When named outputs are used within a Mapper implementation,
+ key/values written to a name output are not part of the reduce phase, only
+ key/values written to the job OutputCollector are part of the
+ reduce phase.
+
+ MultipleOutputs supports counters, by default the are disabled. The counters + group is the {@link MultipleOutputs} class name. +
+ The names of the counters are the same as the named outputs. For multi + named outputs the name of the counter is the concatenation of the named + output, and underscore '_' and the multiname. ++ Job configuration usage pattern is: +
+ + JobConf conf = new JobConf(); + + conf.setInputPath(inDir); + FileOutputFormat.setOutputPath(conf, outDir); + + conf.setMapperClass(MOMap.class); + conf.setReducerClass(MOReduce.class); + ... + + // Defines additional single text based output 'text' for the job + MultipleOutputs.addNamedOutput(conf, "text", TextOutputFormat.class, + LongWritable.class, Text.class); + + // Defines additional multi sequencefile based output 'sequence' for the + // job + MultipleOutputs.addMultiNamedOutput(conf, "seq", + SequenceFileOutputFormat.class, + LongWritable.class, Text.class); + ... + + JobClient jc = new JobClient(); + RunningJob job = jc.submitJob(conf); + + ... ++
+ Job configuration usage pattern is: +
+
+ public class MOReduce implements
+ Reducer<WritableComparable, Writable> {
+ private MultipleOutputs mos;
+
+ public void configure(JobConf conf) {
+ ...
+ mos = new MultipleOutputs(conf);
+ }
+
+ public void reduce(WritableComparable key, Iterator<Writable> values,
+ OutputCollector output, Reporter reporter)
+ throws IOException {
+ ...
+ mos.getCollector("text", reporter).collect(key, new Text("Hello"));
+ mos.getCollector("seq", "A", reporter).collect(key, new Text("Bye"));
+ mos.getCollector("seq", "B", reporter).collect(key, new Text("Chau"));
+ ...
+ }
+
+ public void close() throws IOException {
+ mos.close();
+ ...
+ }
+
+ }
+ ]]>
+ + Map implementations using this MapRunnable must be thread-safe. +
+ The Map-Reduce job has to be configured to use this MapRunnable class (using
+ the JobConf.setMapRunnerClass method) and
+ the number of threads the thread-pool can use with the
+ mapred.map.multithreadedrunner.threads property, its default
+ value is 10 threads.
+
]]> +
ClusterMetrics provides clients with information such as:
+ Clients can query for the latest ClusterMetrics, via
+ {@link Cluster#getClusterStatus()}.
Counters represent global counters, defined either by the
+ Map-Reduce framework or applications. Each Counter is named by
+ an {@link Enum} and has a long for the value.
+
+ Counters are bunched into Groups, each comprising of
+ counters from a particular Enum class.]]>
+
Counters holds per job/task counters, defined either by the
+ Map-Reduce framework or applications. Each Counter can be of
+ any {@link Enum} type.
+
+ Counters are bunched into {@link CounterGroup}s, each
+ comprising of counters from a particular Enum class.]]>
+
Note: The split is a logical split of the inputs and the + input files are not physically split into chunks. For e.g. a split could + be <input-file-path, start, offset> tuple. The InputFormat + also creates the {@link RecordReader} to read the {@link InputSplit}. + + @param context job configuration. + @return an array of {@link InputSplit}s for the job.]]> +
The Map-Reduce framework relies on the InputFormat of the
+ job to:
+
InputSplit for processing by
+ the {@link Mapper}.
+ The default behavior of file-based {@link InputFormat}s, typically + sub-classes of {@link FileInputFormat}, is to split the + input into logical {@link InputSplit}s based on the total size, in + bytes, of the input files. However, the {@link FileSystem} blocksize of + the input files is treated as an upper bound for input splits. A lower bound + on the split size can be set via + + mapreduce.input.fileinputformat.split.minsize.
+ +Clearly, logical splits based on input-size is insufficient for many
+ applications since record boundaries are to respected. In such cases, the
+ application has to also implement a {@link RecordReader} on whom lies the
+ responsibility to respect record-boundaries and present a record-oriented
+ view of the logical InputSplit to the individual task.
+
+ @see InputSplit
+ @see RecordReader
+ @see FileInputFormat]]>
+
Typically, it presents a byte-oriented view on the input and is the + responsibility of {@link RecordReader} of the job to process this and present + a record-oriented view. + + @see InputFormat + @see RecordReader]]> +
Configuration so
+ that any necessary internal modifications do not reflect on the incoming
+ parameter.
+
+ A Cluster will be created from the conf parameter only when it's needed.
+
+ @param conf the configuration
+ @return the {@link Job} , with no connection to a cluster yet.
+ @throws IOException]]>
+ Configuration so
+ that any necessary internal modifications do not reflect on the incoming
+ parameter.
+
+ @param conf the configuration
+ @return the {@link Job} , with no connection to a cluster yet.
+ @throws IOException]]>
+ Configuration so
+ that any necessary internal modifications do not reflect on the incoming
+ parameter.
+
+ @param status job status
+ @param conf job configuration
+ @return the {@link Job} , with no connection to a cluster yet.
+ @throws IOException]]>
+ Configuration so
+ that any necessary internal modifications do not reflect on the incoming
+ parameter.
+
+ @param ignored
+ @return the {@link Job} , with no connection to a cluster yet.
+ @throws IOException
+ @deprecated Use {@link #getInstance()}]]>
+ Configuration so
+ that any necessary internal modifications do not reflect on the incoming
+ parameter.
+
+ @param ignored
+ @param conf job configuration
+ @return the {@link Job} , with no connection to a cluster yet.
+ @throws IOException
+ @deprecated Use {@link #getInstance(Configuration)}]]>
+ false.
+ @throws IOException]]>
+ false.
+ @throws IOException]]>
+ false.]]>
+ false.]]>
+ false.]]>
+ + Normally the user creates the application, describes various facets of the + job via {@link Job} and then submits the job and monitor its progress.
+ +Here is an example on how to submit a job:
+
+ // Create a new Job
+ Job job = Job.getInstance();
+ job.setJarByClass(MyJob.class);
+
+ // Specify various job-specific parameters
+ job.setJobName("myjob");
+
+ job.setInputPath(new Path("in"));
+ job.setOutputPath(new Path("out"));
+
+ job.setMapperClass(MyJob.MyMapper.class);
+ job.setReducerClass(MyJob.MyReducer.class);
+
+ // Submit the job, then poll for progress until the job is complete
+ job.waitForCompletion(true);
+ job_200707121733_0003 , which represents the third job
+ running at the jobtracker started at 200707121733.
+ + Applications should never construct or parse JobID strings, but rather + use appropriate constructors or {@link #forName(String)} method. + + @see TaskID + @see TaskAttemptID]]> +
The Hadoop Map-Reduce framework spawns one map task for each
+ {@link InputSplit} generated by the {@link InputFormat} for the job.
+ Mapper implementations can access the {@link Configuration} for
+ the job via the {@link JobContext#getConfiguration()}.
+
+
The framework first calls
+ {@link #setup(org.apache.hadoop.mapreduce.Mapper.Context)}, followed by
+ {@link #map(Object, Object, org.apache.hadoop.mapreduce.Mapper.Context)}
+ for each key/value pair in the InputSplit. Finally
+ {@link #cleanup(org.apache.hadoop.mapreduce.Mapper.Context)} is called.
All intermediate values associated with a given output key are + subsequently grouped by the framework, and passed to a {@link Reducer} to + determine the final output. Users can control the sorting and grouping by + specifying two key {@link RawComparator} classes.
+ +The Mapper outputs are partitioned per
+ Reducer. Users can control which keys (and hence records) go to
+ which Reducer by implementing a custom {@link Partitioner}.
+
+
Users can optionally specify a combiner, via
+ {@link Job#setCombinerClass(Class)}, to perform local aggregation of the
+ intermediate outputs, which helps to cut down the amount of data transferred
+ from the Mapper to the Reducer.
+
+
Applications can specify if and how the intermediate
+ outputs are to be compressed and which {@link CompressionCodec}s are to be
+ used via the Configuration.
If the job has zero
+ reduces then the output of the Mapper is directly written
+ to the {@link OutputFormat} without sorting by keys.
Example:
+
+ public class TokenCounterMapper
+ extends Mapper<Object, Text, Text, IntWritable>{
+
+ private final static IntWritable one = new IntWritable(1);
+ private Text word = new Text();
+
+ public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
+ StringTokenizer itr = new StringTokenizer(value.toString());
+ while (itr.hasMoreTokens()) {
+ word.set(itr.nextToken());
+ context.write(word, one);
+ }
+ }
+ }
+ Applications may override the
+ {@link #run(org.apache.hadoop.mapreduce.Mapper.Context)} method to exert
+ greater control on map processing e.g. multi-threaded Mappers
+ etc.
false otherwise
+ @see #recoverTask(TaskAttemptContext)
+ @deprecated Use {@link #isRecoverySupported(JobContext)} instead.]]>
+ false otherwise
+ @throws IOException]]>
+ false otherwise
+ @throws IOException
+ @see #recoverTask(TaskAttemptContext)]]>
+ The Map-Reduce framework relies on the OutputCommitter of
+ the job to:
+
The Map-Reduce framework relies on the OutputFormat of the
+ job to:
+
key.]]>
+ Partitioner controls the partitioning of the keys of the
+ intermediate map-outputs. The key (or a subset of the key) is used to derive
+ the partition, typically by a hash function. The total number of partitions
+ is the same as the number of reduce tasks for the job. Hence this controls
+ which of the m reduce tasks the intermediate key (and hence the
+ record) is sent for reduction.
+
+ Note: A Partitioner is created only when there are multiple
+ reducers.
Note: If you require your Partitioner class to obtain the Job's + configuration object, implement the {@link Configurable} interface.
+ + @see Reducer]]> +RecordWriter implementations write the job outputs to the
+ {@link FileSystem}.
+
+ @see OutputFormat]]>
+
Reducer implementations
+ can access the {@link Configuration} for the job via the
+ {@link JobContext#getConfiguration()} method.
+
+ Reducer has 3 primary phases:
The Reducer copies the sorted output from each
+ {@link Mapper} using HTTP across the network.
The framework merge sorts Reducer inputs by
+ keys
+ (since different Mappers may have output the same key).
The shuffle and sort phases occur simultaneously i.e. while outputs are + being fetched they are merged.
+ + SecondarySort + +To achieve a secondary sort on the values returned by the value + iterator, the application should extend the key with the secondary + key and define a grouping comparator. The keys will be sorted using the + entire key, but will be grouped using the grouping comparator to decide + which keys and values are sent in the same call to reduce.The grouping + comparator is specified via + {@link Job#setGroupingComparatorClass(Class)}. The sort order is + controlled by + {@link Job#setSortComparatorClass(Class)}.
+ + + For example, say that you want to find duplicate web pages and tag them + all with the url of the "best" known example. You would set up the job + like: +In this phase the
+ {@link #reduce(Object, Iterable, org.apache.hadoop.mapreduce.Reducer.Context)}
+ method is called for each <key, (collection of values)> in
+ the sorted inputs.
The output of the reduce task is typically written to a + {@link RecordWriter} via + {@link Context#write(Object, Object)}.
+The output of the Reducer is not re-sorted.
Example:
+
+ public class IntSumReducer<Key> extends Reducer<Key,IntWritable,
+ Key,IntWritable> {
+ private IntWritable result = new IntWritable();
+
+ public void reduce(Key key, Iterable<IntWritable> values,
+ Context context) throws IOException, InterruptedException {
+ int sum = 0;
+ for (IntWritable val : values) {
+ sum += val.get();
+ }
+ result.set(sum);
+ context.write(key, result);
+ }
+ }
+ Counter for the given counterName]]>
+ counterName.
+ @param counterName counter name
+ @return the Counter for the given groupName and
+ counterName]]>
+ attempt_200707121733_0003_m_000005_0 , which represents the
+ zeroth task attempt for the fifth map task in the third job
+ running at the jobtracker started at 200707121733.
+ + Applications should never construct or parse TaskAttemptID strings + , but rather use appropriate constructors or {@link #forName(String)} + method. + + @see JobID + @see TaskID]]> +
task_200707121733_0003_m_000005 , which represents the
+ fifth map task in the third job running at the jobtracker
+ started at 200707121733.
+ + Applications should never construct or parse TaskID strings + , but rather use appropriate constructors or {@link #forName(String)} + method. + + @see JobID + @see TaskAttemptID]]> +
mapperConf, have precedence over the job's Configuration. This
+ precedence is in effect when the task is running.
+
+ + IMPORTANT: There is no need to specify the output key/value classes for the + ChainMapper, this is done by the addMapper for the last mapper in the chain +
+ + @param job + The job. + @param klass + the Mapper class to add. + @param inputKeyClass + mapper input key class. + @param inputValueClass + mapper input value class. + @param outputKeyClass + mapper output key class. + @param outputValueClass + mapper output value class. + @param mapperConf + a configuration for the Mapper class. It is recommended to use a + Configuration without default values using the +Configuration(boolean loadDefaults) constructor with
+ FALSE.]]>
+ + The key functionality of this feature is that the Mappers in the chain do not + need to be aware that they are executed in a chain. This enables having + reusable specialized Mappers that can be combined to perform composite + operations within a single task. +
++ Special care has to be taken when creating chains that the key/values output + by a Mapper are valid for the following Mapper in the chain. It is assumed + all Mappers and the Reduce in the chain use matching output and input key and + value classes as no conversion is done by the chaining code. +
+
+ Using the ChainMapper and the ChainReducer classes is possible to compose
+ Map/Reduce jobs that look like [MAP+ / REDUCE MAP*]. And
+ immediate benefit of this pattern is a dramatic reduction in disk IO.
+
+ IMPORTANT: There is no need to specify the output key/value classes for the + ChainMapper, this is done by the addMapper for the last mapper in the chain. +
+ ChainMapper usage pattern: ++ +
+ ... + Job = new Job(conf); + + Configuration mapAConf = new Configuration(false); + ... + ChainMapper.addMapper(job, AMap.class, LongWritable.class, Text.class, + Text.class, Text.class, true, mapAConf); + + Configuration mapBConf = new Configuration(false); + ... + ChainMapper.addMapper(job, BMap.class, Text.class, Text.class, + LongWritable.class, Text.class, false, mapBConf); + + ... + + job.waitForComplettion(true); + ... +]]> +
reducerConf, have precedence over the job's Configuration.
+ This precedence is in effect when the task is running.
+
+ + IMPORTANT: There is no need to specify the output key/value classes for the + ChainReducer, this is done by the setReducer or the addMapper for the last + element in the chain. +
+ + @param job + the job + @param klass + the Reducer class to add. + @param inputKeyClass + reducer input key class. + @param inputValueClass + reducer input value class. + @param outputKeyClass + reducer output key class. + @param outputValueClass + reducer output value class. + @param reducerConf + a configuration for the Reducer class. It is recommended to use a + Configuration without default values using the +Configuration(boolean loadDefaults) constructor with
+ FALSE.]]>
+ mapperConf, have precedence over the job's Configuration. This
+ precedence is in effect when the task is running.
+
+ + IMPORTANT: There is no need to specify the output key/value classes for the + ChainMapper, this is done by the addMapper for the last mapper in the + chain. +
+ + @param job + The job. + @param klass + the Mapper class to add. + @param inputKeyClass + mapper input key class. + @param inputValueClass + mapper input value class. + @param outputKeyClass + mapper output key class. + @param outputValueClass + mapper output value class. + @param mapperConf + a configuration for the Mapper class. It is recommended to use a + Configuration without default values using the +Configuration(boolean loadDefaults) constructor with
+ FALSE.]]>
+ + The key functionality of this feature is that the Mappers in the chain do not + need to be aware that they are executed after the Reducer or in a chain. This + enables having reusable specialized Mappers that can be combined to perform + composite operations within a single task. +
++ Special care has to be taken when creating chains that the key/values output + by a Mapper are valid for the following Mapper in the chain. It is assumed + all Mappers and the Reduce in the chain use matching output and input key and + value classes as no conversion is done by the chaining code. +
+ Using the ChainMapper and the ChainReducer classes is possible to
+ compose Map/Reduce jobs that look like [MAP+ / REDUCE MAP*]. And
+ immediate benefit of this pattern is a dramatic reduction in disk IO.
+ IMPORTANT: There is no need to specify the output key/value classes for the + ChainReducer, this is done by the setReducer or the addMapper for the last + element in the chain. +
+ ChainReducer usage pattern: ++ +
+ ... + Job = new Job(conf); + .... + + Configuration reduceConf = new Configuration(false); + ... + ChainReducer.setReducer(job, XReduce.class, LongWritable.class, Text.class, + Text.class, Text.class, true, reduceConf); + + ChainReducer.addMapper(job, CMap.class, Text.class, Text.class, + LongWritable.class, Text.class, false, null); + + ChainReducer.addMapper(job, DMap.class, LongWritable.class, Text.class, + LongWritable.class, LongWritable.class, true, null); + + ... + + job.waitForCompletion(true); + ... +]]> +
+ Implementations are responsible for writing the fields of the object + to PreparedStatement, and reading the fields of the object from the + ResultSet. + +
Example:
+ If we have the following table in the database : ++ CREATE TABLE MyTable ( + counter INTEGER NOT NULL, + timestamp BIGINT NOT NULL, + ); ++ then we can read/write the tuples from/to the table with : +
+ public class MyWritable implements Writable, DBWritable {
+ // Some data
+ private int counter;
+ private long timestamp;
+
+ //Writable#write() implementation
+ public void write(DataOutput out) throws IOException {
+ out.writeInt(counter);
+ out.writeLong(timestamp);
+ }
+
+ //Writable#readFields() implementation
+ public void readFields(DataInput in) throws IOException {
+ counter = in.readInt();
+ timestamp = in.readLong();
+ }
+
+ public void write(PreparedStatement statement) throws SQLException {
+ statement.setInt(1, counter);
+ statement.setLong(2, timestamp);
+ }
+
+ public void readFields(ResultSet resultSet) throws SQLException {
+ counter = resultSet.getInt(1);
+ timestamp = resultSet.getLong(2);
+ }
+ }
+ ]]>
+ CombineFileSplit's.
+
+ @see CombineFileSplit]]>
+ SequenceFileInputFormat.
+
+ @see CombineFileInputFormat]]>
+ TextInputFormat.
+
+ @see CombineFileInputFormat]]>
+ FileInputFormat implementations can override this and return
+ false to ensure that individual input files are never split-up
+ so that {@link Mapper}s process entire files.
+
+ @param context the job context
+ @param filename the file name to check
+ @return is this file splitable?]]>
+ FileInputFormat is the base class for all file-based
+ InputFormats. This provides a generic implementation of
+ {@link #getSplits(JobContext)}.
+
+ Implementations of FileInputFormat can also override the
+ {@link #isSplitable(JobContext, Path)} method to prevent input files
+ from being split-up in certain situations. Implementations that may
+ deal with non-splittable files must override this method, since
+ the default implementation assumes splitting is always possible.]]>
+ ) }]]> +
+ Mapper implementations using this MapRunnable must be thread-safe. +
+ The Map-Reduce job has to be configured with the mapper to use via + {@link #setMapperClass(Job, Class)} and + the number of thread the thread-pool can use with the + {@link #getNumberOfThreads(JobContext)} method. The default + value is 10 threads. +
]]> +
Mapper.Context for custom implementations]]>
+ false otherwise]]>
+ Some applications need to create/write-to side-files, which differ from + the actual job-outputs. + +
In such cases there could be issues with 2 instances of the same TIP + (running simultaneously e.g. speculative tasks) trying to open/write-to the + same file (path) on HDFS. Hence the application-writer will have to pick + unique names per task-attempt (e.g. using the attemptid, say + attempt_200709221812_0001_m_000000_0), not just per TIP.
+ +To get around this the Map-Reduce framework helps the application-writer + out by maintaining a special + ${mapreduce.output.fileoutputformat.outputdir}/_temporary/_${taskid} + sub-directory for each task-attempt on HDFS where the output of the + task-attempt goes. On successful completion of the task-attempt the files + in the ${mapreduce.output.fileoutputformat.outputdir}/_temporary/_${taskid} (only) + are promoted to ${mapreduce.output.fileoutputformat.outputdir}. Of course, the + framework discards the sub-directory of unsuccessful task-attempts. This + is completely transparent to the application.
+ +The application-writer can take advantage of this by creating any + side-files required in a work directory during execution + of his task i.e. via + {@link #getWorkOutputPath(TaskInputOutputContext)}, and + the framework will move them out similarly - thus she doesn't have to pick + unique paths per task-attempt.
+ +The entire discussion holds true for maps of jobs with + reducer=NONE (i.e. 0 reduces) since output of the map, in that case, + goes directly to HDFS.
+ + @return the {@link Path} to the task's temporary output directory + for the map-reduce job.]]> +This method uses the {@link #getUniqueFile} method to make the file name + unique for the task.
+ + @param context the context for the task. + @param name the name for the file. + @param extension the extension for the file + @return a unique path accross all tasks of the job.]]> +close()]]>
+ OutputFormat, with its own key class and with its own value
+ class.
+
+
+ + Case two: to write data to different files provided by user +
+ ++ MultipleOutputs supports counters, by default they are disabled. The + counters group is the {@link MultipleOutputs} class name. The names of the + counters are the same as the output name. These count the number records + written to each output name. +
+ + Usage pattern for job submission: ++ + Job job = new Job(); + + FileInputFormat.setInputPath(job, inDir); + FileOutputFormat.setOutputPath(job, outDir); + + job.setMapperClass(MOMap.class); + job.setReducerClass(MOReduce.class); + ... + + // Defines additional single text based output 'text' for the job + MultipleOutputs.addNamedOutput(job, "text", TextOutputFormat.class, + LongWritable.class, Text.class); + + // Defines additional sequence-file based output 'sequence' for the job + MultipleOutputs.addNamedOutput(job, "seq", + SequenceFileOutputFormat.class, + LongWritable.class, Text.class); + ... + + job.waitForCompletion(true); + ... ++
+ Usage in Reducer: +
+ <K, V> String generateFileName(K k, V v) {
+ return k.toString() + "_" + v.toString();
+ }
+
+ public class MOReduce extends
+ Reducer<WritableComparable, Writable,WritableComparable, Writable> {
+ private MultipleOutputs mos;
+ public void setup(Context context) {
+ ...
+ mos = new MultipleOutputs(context);
+ }
+
+ public void reduce(WritableComparable key, Iterator<Writable> values,
+ Context context)
+ throws IOException {
+ ...
+ mos.write("text", , key, new Text("Hello"));
+ mos.write("seq", LongWritable(1), new Text("Bye"), "seq_a");
+ mos.write("seq", LongWritable(2), key, new Text("Chau"), "seq_b");
+ mos.write(key, new Text("value"), generateFileName(key, new Text("value")));
+ ...
+ }
+
+ public void cleanup(Context) throws IOException {
+ mos.close();
+ ...
+ }
+
+ }
+
+
+ + When used in conjuction with org.apache.hadoop.mapreduce.lib.output.LazyOutputFormat, + MultipleOutputs can mimic the behaviour of MultipleTextOutputFormat and MultipleSequenceFileOutputFormat + from the old Hadoop API - ie, output can be written from the Reducer to more than one location. +
+ +
+ Use MultipleOutputs.write(KEYOUT key, VALUEOUT value, String baseOutputPath) to write key and
+ value to a path specified by baseOutputPath, with no need to specify a named output.
+ Warning: when the baseOutputPath passed to MultipleOutputs.write
+ is a path that resolves outside of the final job output directory, the
+ directory is created immediately and then persists through subsequent
+ task retries, breaking the concept of output committing:
+
+ private MultipleOutputs<Text, Text> out;
+
+ public void setup(Context context) {
+ out = new MultipleOutputs<Text, Text>(context);
+ ...
+ }
+
+ public void reduce(Text key, Iterable<Text> values, Context context) throws IOException, InterruptedException {
+ for (Text t : values) {
+ out.write(key, t, generateFileName(<parameter list...>));
+ }
+ }
+
+ protected void cleanup(Context context) throws IOException, InterruptedException {
+ out.close();
+ }
+
+
+
+ Use your own code in generateFileName() to create a custom path to your results.
+ '/' characters in baseOutputPath will be translated into directory levels in your file system.
+ Also, append your custom-generated path with "part" or similar, otherwise your output will be -00000, -00001 etc.
+ No call to context.write() is necessary. See example generateFileName() code below.
+
+ private String generateFileName(Text k) {
+ // expect Text k in format "Surname|Forename"
+ String[] kStr = k.toString().split("\\|");
+
+ String sName = kStr[0];
+ String fName = kStr[1];
+
+ // example for k = Smith|John
+ // output written to /user/hadoop/path/to/output/Smith/John-r-00000 (etc)
+ return sName + "/" + fName;
+ }
+
+
+
+ Using MultipleOutputs in this way will still create zero-sized default output, eg part-00000.
+ To prevent this use LazyOutputFormat.setOutputFormatClass(job, TextOutputFormat.class);
+ instead of job.setOutputFormatClass(TextOutputFormat.class); in your Hadoop job configuration.
+
The subarray to be used for the partitioning can be defined by means + of the following properties: +
+ +---+---+---+---+---+
+ | B | B | B | B | B |
+ +---+---+---+---+---+
+ 0 1 2 3 4
+ -5 -4 -3 -2 -1
+ i
+ and j as first and last element, repectively, when
+ i and j are the left and right offset.
+
+ For Hadoop programs written in Java, it is advisable to use one of + the following static convenience methods for setting the offsets: +
Reducer.Context for custom implementations]]>
+ YARN verifies access to the application, queue
+ etc. before accepting the request.
+
+ If the user does not have VIEW_APP access then the following
+ fields in the report will be set to stubbed values:
+
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ In secure mode, YARN verifies access to the application, queue
+ etc. before accepting the request.
+
+ In secure mode, YARN verifies access to the application, queue
+ etc. before accepting the request.
+
ResourceManager. New containers assigned to the master are
+ retrieved. Status of completed containers and node health updates are also
+ retrieved. This also doubles up as a heartbeat to the ResourceManager and
+ must be made periodically. The call may not always return any new
+ allocations of containers. App should not make concurrent allocate
+ requests. May cause request loss.
+
+ + Note : If the user has not removed container requests that have already + been satisfied, then the re-register may end up sending the entire + container requests to the RM (including matched requests). Which would mean + the RM could end up giving it a lot of new allocated containers. +
+ + @param progressIndicator Indicates progress made by the master + @return the response of the allocate request + @throws YarnException + @throws IOException]]> +Containers that are returned from previous successful
+ allocations or resource changes. By passing in the existing container and a
+ target resource capability to this method, the application requests the
+ ResourceManager to change the existing resource allocation to the target
+ resource allocation.
+
+ @deprecated use
+ {@link #requestContainerUpdate(Container, UpdateContainerRequest)}
+
+ @param container The container returned from the last successful resource
+ allocation or resource change
+ @param capability The target resource capability of the container]]>
+ UpdateContainerRequest.]]>
+ addContainerRequest earlier in the lifecycle. For performance,
+ the AMRMClient may return its internal collection directly without creating
+ a copy. Users should not perform mutable operations on the return value.
+ Each collection in the list contains requests with identical
+ Resource size that fit in the given capability. In a
+ collection, requests will be returned in the same order as they were added.
+
+ NOTE: This API only matches Container requests that were created by the
+ client WITHOUT the allocationRequestId being set.
+
+ @return Collection of request matching the parameters]]>
+ addContainerRequest earlier in the lifecycle. For performance,
+ the AMRMClient may return its internal collection directly without creating
+ a copy. Users should not perform mutable operations on the return value.
+ Each collection in the list contains requests with identical
+ Resource size that fit in the given capability. In a
+ collection, requests will be returned in the same order as they were added.
+ specify an ExecutionType.
+
+ NOTE: This API only matches Container requests that were created by the
+ client WITHOUT the allocationRequestId being set.
+
+ @param priority Priority
+ @param resourceName Location
+ @param executionType ExecutionType
+ @param capability Capability
+ @return Collection of request matching the parameters]]>
+ addContainerRequest earlier in the lifecycle. For performance,
+ the AMRMClient may return its internal collection directly without creating
+ a copy. Users should not perform mutable operations on the return value.
+
+ NOTE: This API only matches Container requests that were created by the
+ client WITH the allocationRequestId being set to a non-default value.
+
+ @param allocationRequestId Allocation Request Id
+ @return Collection of request matching the parameters]]>
+ AMRMClient
+ + If a NM token cache is not set, the {@link NMTokenCache#getSingleton()} + singleton instance will be used. + + @param nmTokenCache the NM token cache to use.]]> +
AMRMClient.
+ + If a NM token cache is not set, the {@link NMTokenCache#getSingleton()} + singleton instance will be used. + + @return the NM token cache.]]> +
checkEveryMillis ms.
+ See also {@link #waitFor(java.util.function.Supplier, int, int)}
+ @param check user defined checker
+ @param checkEveryMillis interval to call check]]>
+ checkEveryMillis ms. In the main loop, this method will log
+ the message "waiting in main loop" for each logInterval times
+ iteration to confirm the thread is alive.
+ @param check user defined checker
+ @param checkEveryMillis interval to call check
+ @param logInterval interval to log for each]]>
+ The ApplicationMaster or other applications that use the
+ client must provide the details of the allocated container, including the
+ Id, the assigned node's Id and the token via {@link Container}. In
+ addition, the AM needs to provide the {@link ContainerLaunchContext} as
+ well.
NodeManager to launch the
+ container
+ @return a map between the auxiliary service names and their outputs
+ @throws YarnException YarnException.
+ @throws IOException IOException.]]>
+ The ApplicationMaster or other applications that use the
+ client must provide the details of the container, including the Id and
+ the target resource encapsulated in the updated container token via
+ {@link Container}.
+
The ApplicationMaster or other applications that use the
+ client must provide the details of the container, including the Id and
+ the target resource encapsulated in the updated container token via
+ {@link Container}.
+
NodeManager
+
+ @throws YarnException YarnException.
+ @throws IOException IOException.]]>
+ NodeManager
+
+ @return the status of a container.
+
+ @throws YarnException YarnException.
+ @throws IOException IOException.]]>
+ NMClient
+ + If a NM token cache is not set, the {@link NMTokenCache#getSingleton()} + singleton instance will be used. + + @param nmTokenCache the NM token cache to use.]]> +
NMClient
+ + If a NM token cache is not set, the {@link NMTokenCache#getSingleton()} + singleton instance will be used. + + @return the NM token cache]]> +
+ NMTokenCache nmTokenCache = new NMTokenCache(); + AMRMClient rmClient = AMRMClient.createAMRMClient(); + NMClient nmClient = NMClient.createNMClient(); + nmClient.setNMTokenCache(nmTokenCache); + ... ++
+ NMTokenCache nmTokenCache = new NMTokenCache();
+ AMRMClient rmClient = AMRMClient.createAMRMClient();
+ NMClient nmClient = NMClient.createNMClient();
+ nmClient.setNMTokenCache(nmTokenCache);
+ AMRMClientAsync rmClientAsync = new AMRMClientAsync(rmClient, 1000, [AMRM_CALLBACK]);
+ NMClientAsync nmClientAsync = new NMClientAsync("nmClient", nmClient, [NM_CALLBACK]);
+ ...
+
+
+ NMTokenCache nmTokenCache = new NMTokenCache();
+ ...
+ ApplicationMasterProtocol amPro = ClientRMProxy.createRMProxy(conf, ApplicationMasterProtocol.class);
+ ...
+ AllocateRequest allocateRequest = ...
+ ...
+ AllocateResponse allocateResponse = rmClient.allocate(allocateRequest);
+ for (NMToken token : allocateResponse.getNMTokens()) {
+ nmTokenCache.setToken(token.getNodeId().toString(), token.getToken());
+ }
+ ...
+ ContainerManagementProtocolProxy nmPro = ContainerManagementProtocolProxy(conf, nmTokenCache);
+ ...
+ nmPro.startContainer(container, containerContext);
+ ...
+
+ SharedCacheManager.
+ The client uses a checksum to identify the resource and an
+ {@link ApplicationId} to identify which application will be using the
+ resource.
+
+
+
+ The SharedCacheManager responds with whether or not the
+ resource exists in the cache. If the resource exists, a URL to
+ the resource in the shared cache is returned. If the resource does not
+ exist, null is returned instead.
+
+ Once a URL has been returned for a resource, that URL is safe to use for + the lifetime of the application that corresponds to the provided + ApplicationId. +
+ + @param applicationId ApplicationId of the application using the resource + @param resourceKey the key (i.e. checksum) that identifies the resource + @return URL to the resource, or null if it does not exist]]> +SharedCacheManager.
+ This method is called once an application is no longer using a claimed
+ resource in the shared cache. The client uses a checksum to identify the
+ resource and an {@link ApplicationId} to identify which application is
+ releasing the resource.
+
+
+ + Note: This method is an optimization and the client is not required to call + it for correctness. +
+ + @param applicationId ApplicationId of the application releasing the + resource + @param resourceKey the key (i.e. checksum) that identifies the resource]]> +YARN. It is a blocking call - it
+ will not return {@link ApplicationId} until the submitted application is
+ submitted successfully and accepted by the ResourceManager.
+
+
+ + Users should provide an {@link ApplicationId} as part of the parameter + {@link ApplicationSubmissionContext} when submitting a new application, + otherwise it will throw the {@link ApplicationIdNotProvidedException}. +
+ +This internally calls {@link ApplicationClientProtocol#submitApplication + (SubmitApplicationRequest)}, and after that, it internally invokes + {@link ApplicationClientProtocol#getApplicationReport + (GetApplicationReportRequest)} and waits till it can make sure that the + application gets properly submitted. If RM fails over or RM restart + happens before ResourceManager saves the application's state, + {@link ApplicationClientProtocol + #getApplicationReport(GetApplicationReportRequest)} will throw + the {@link ApplicationNotFoundException}. This API automatically resubmits + the application with the same {@link ApplicationSubmissionContext} when it + catches the {@link ApplicationNotFoundException}
+ + @param appContext + {@link ApplicationSubmissionContext} containing all the details + needed to submit a new application + @return {@link ApplicationId} of the accepted application + @throws YarnException + @throws IOException + @see #createApplication()]]> +
+ In secure mode, YARN verifies access to the application, queue
+ etc. before accepting the request.
+
+ If the user does not have VIEW_APP access then the following
+ fields in the report will be set to stubbed values:
+
+ The AMRM token will be returned only if all the following conditions are + met: +
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ If the user does not have VIEW_APP access for an application
+ then the corresponding report will be filtered as described in
+ {@link #getApplicationReport(ApplicationId)}.
+
+ In secure mode, YARN verifies access to the application, queue
+ etc. before accepting the request.
+
+ In secure mode, YARN verifies access to the application, queue
+ etc. before accepting the request.
+
+ The client packages all details of its request in a + {@link ReservationSubmissionRequest} object. This contains information + about the amount of capacity, temporal constraints, and gang needs. + Furthermore, the reservation might be composed of multiple stages, with + ordering dependencies among them. +
+ ++ In order to respond, a new admission control component in the + {@code ResourceManager} performs an analysis of the resources that have + been committed over the period of time the user is requesting, verify that + the user requests can be fulfilled, and that it respect a sharing policy + (e.g., {@code CapacityOverTimePolicy}). Once it has positively determined + that the ReservationRequest is satisfiable the {@code ResourceManager} + answers with a {@link ReservationSubmissionResponse} that includes a + {@link ReservationId}. Upon failure to find a valid allocation the response + is an exception with the message detailing the reason of failure. +
+ ++ The semantics guarantees that the {@link ReservationId} returned, + corresponds to a valid reservation existing in the time-range request by + the user. The amount of capacity dedicated to such reservation can vary + overtime, depending of the allocation that has been determined. But it is + guaranteed to satisfy all the constraint expressed by the user in the + {@link ReservationDefinition} +
+ + @param request request to submit a new Reservation + @return response contains the {@link ReservationId} on accepting the + submission + @throws YarnException if the reservation cannot be created successfully + @throws IOException]]> ++ The allocation is attempted by virtually substituting all previous + allocations related to this Reservation with new ones, that satisfy the new + {@link ReservationDefinition}. Upon success the previous allocation is + atomically substituted by the new one, and on failure (i.e., if the system + cannot find a valid allocation for the updated request), the previous + allocation remains valid. +
+ + @param request to update an existing Reservation (the + {@link ReservationUpdateRequest} should refer to an existing valid + {@link ReservationId}) + @return response empty on successfully updating the existing reservation + @throws YarnException if the request is invalid or reservation cannot be + updated successfully + @throws IOException]]> +ResourceManager]]>
+ ResourceManager]]>
+ Containers that are returned from previous successful
+ allocations or resource changes. By passing in the existing container and a
+ target resource capability to this method, the application requests the
+ ResourceManager to change the existing resource allocation to the target
+ resource allocation.
+
+ @deprecated use
+ {@link #requestContainerUpdate(Container, UpdateContainerRequest)}
+
+ @param container The container returned from the last successful resource
+ allocation or resource change
+ @param capability The target resource capability of the container]]>
+ UpdateContainerRequest.]]>
+ checkEveryMillis ms.
+ See also {@link #waitFor(java.util.function.Supplier, int, int)}
+ @param check user defined checker
+ @param checkEveryMillis interval to call check]]>
+ checkEveryMillis ms. In the main loop, this method will log
+ the message "waiting in main loop" for each logInterval times
+ iteration to confirm the thread is alive.
+ @param check user defined checker
+ @param checkEveryMillis interval to call check
+ @param logInterval interval to log for each]]>
+
+ {@code
+ class MyCallbackHandler extends AMRMClientAsync.AbstractCallbackHandler {
+ public void onContainersAllocated(List containers) {
+ [run tasks on the containers]
+ }
+
+ public void onContainersUpdated(List containers) {
+ [determine if resource allocation of containers have been increased in
+ the ResourceManager, and if so, inform the NodeManagers to increase the
+ resource monitor/enforcement on the containers]
+ }
+
+ public void onContainersCompleted(List statuses) {
+ [update progress, check whether app is done]
+ }
+
+ public void onNodesUpdated(List updated) {}
+
+ public void onReboot() {}
+ }
+ }
+
+
+ The client's lifecycle should be managed similarly to the following:
+
+
+ {@code
+ AMRMClientAsync asyncClient =
+ createAMRMClientAsync(appAttId, 1000, new MyCallbackhandler());
+ asyncClient.init(conf);
+ asyncClient.start();
+ RegisterApplicationMasterResponse response = asyncClient
+ .registerApplicationMaster(appMasterHostname, appMasterRpcPort,
+ appMasterTrackingUrl);
+ asyncClient.addContainerRequest(containerRequest);
+ [... wait for application to complete]
+ asyncClient.unregisterApplicationMaster(status, appMsg, trackingUrl);
+ asyncClient.stop();
+ }
+ ]]>
+ The ApplicationMaster or other applications that use the
+ client must provide the details of the container, including the Id and
+ the target resource encapsulated in the updated container token via
+ {@link Container}.
+
+ {@code
+ class MyCallbackHandler extends NMClientAsync.AbstractCallbackHandler {
+ public void onContainerStarted(ContainerId containerId,
+ Map allServiceResponse) {
+ [post process after the container is started, process the response]
+ }
+
+ public void onContainerResourceIncreased(ContainerId containerId,
+ Resource resource) {
+ [post process after the container resource is increased]
+ }
+
+ public void onContainerStatusReceived(ContainerId containerId,
+ ContainerStatus containerStatus) {
+ [make use of the status of the container]
+ }
+
+ public void onContainerStopped(ContainerId containerId) {
+ [post process after the container is stopped]
+ }
+
+ public void onStartContainerError(
+ ContainerId containerId, Throwable t) {
+ [handle the raised exception]
+ }
+
+ public void onGetContainerStatusError(
+ ContainerId containerId, Throwable t) {
+ [handle the raised exception]
+ }
+
+ public void onStopContainerError(
+ ContainerId containerId, Throwable t) {
+ [handle the raised exception]
+ }
+ }
+ }
+
+
+ The client's life-cycle should be managed like the following:
+
+
+ {@code
+ NMClientAsync asyncClient =
+ NMClientAsync.createNMClientAsync(new MyCallbackhandler());
+ asyncClient.init(conf);
+ asyncClient.start();
+ asyncClient.startContainer(container, containerLaunchContext);
+ [... wait for container being started]
+ asyncClient.getContainerStatus(container.getId(), container.getNodeId(),
+ container.getContainerToken());
+ [... handle the status in the callback instance]
+ asyncClient.stopContainer(container.getId(), container.getNodeId(),
+ container.getContainerToken());
+ [... wait for container being stopped]
+ asyncClient.stop();
+ }
+ ]]>
+ yarn.sharedcache.checksum.algo.impl)
+
+ @return SharedCacheChecksum object]]>
+ + Note that header and ellipses are not counted against {@link #limit}. +
+ An example: + +
+ {@code
+ // At the beginning it's an empty string
+ final Appendable shortAppender = new BoundedAppender(80);
+ // The whole message fits into limit
+ shortAppender.append(
+ "message1 this is a very long message but fitting into limit\n");
+ // The first message is truncated, the second not
+ shortAppender.append("message2 this is shorter than the previous one\n");
+ // The first message is deleted, the second truncated, the third
+ // preserved
+ shortAppender.append("message3 this is even shorter message, maybe.\n");
+ // The first two are deleted, the third one truncated, the last preserved
+ shortAppender.append("message4 the shortest one, yet the greatest :)");
+ // Current contents are like this:
+ // Diagnostic messages truncated, showing last 80 chars out of 199:
+ // ...s is even shorter message, maybe.
+ // message4 the shortest one, yet the greatest :)
+ }
+
+ + Note that null values are {@link #append(CharSequence) append}ed + just like in {@link StringBuilder#append(CharSequence) original + implementation}. +
+ Note that this class is not thread safe.]]> +