Package org.h2.tools

Class DirectRecover

java.lang.Object

org.h2.util.Tool

org.h2.tools.Recover

org.h2.tools.DirectRecover

All Implemented Interfaces:

org.h2.store.DataHandler

public class DirectRecover
extends Recover

Enhanced database recovery tool with streaming and compression support.

DirectRecover extends the original Recover class to provide:

• Streaming Processing: Uses pipes for memory-efficient processing of large databases
• Parallel Execution: Dump generation and SQL writing happen concurrently
• Compression Support: On-the-fly compression with GZIP, ZIP, BZIP2, or KANZI
• No Intermediate Files: Direct streaming prevents disk space issues

Usage Examples:

Basic Recovery (No Compression):

 java -cp h2.jar org.h2.tools.DirectRecover -dir /path/to/db -db mydb
 

With Built-in Compression:

 # GZIP compression (recommended for most cases)
 java -cp h2.jar org.h2.tools.DirectRecover -dir /path/to/db -db mydb -compress gzip

 # ZIP compression (widely compatible)
 java -cp h2.jar org.h2.tools.DirectRecover -dir /path/to/db -db mydb -compress zip
 

With Optional External Libraries:

 # BZIP2 compression (requires Apache Commons Compress)
 java -cp "h2.jar:commons-compress.jar" org.h2.tools.DirectRecover -dir /path/to/db -db mydb -compress bzip2

 # KANZI compression (requires Kanzi library, best compression)
 java -cp "h2.jar:kanzi.jar" org.h2.tools.DirectRecover -dir /path/to/db -db mydb -compress kanzi
 

Debug and Troubleshooting:

 # Enable verbose debug output
 java -cp h2.jar org.h2.tools.DirectRecover -dir /path/to/db -db mydb -trace -compress gzip

 # Process all databases in directory
 java -cp h2.jar org.h2.tools.DirectRecover -dir /path/to/db -compress gzip
 

Command Line Options:

Supported Options

Option	Description	Default
-dir <directory>	Database directory	. (current directory)
-db <database>	Database name (without .mv.db extension)	All databases in directory
-compress <type>	Compression: none, gzip, zip, bzip2, kanzi	none
-trace	Enable debug output and verbose logging	false
-help or -?	Show help information	-

Output Files:

• No compression: database.sql
• GZIP: database.sql.gz
• ZIP: database.sql.zip
• BZIP2: database.sql.bz2
• KANZI: database.sql.knz

External Dependencies (Optional):

• Apache Commons Compress: Required for BZIP2 compression
  Download: https://commons.apache.org/proper/commons-compress/
• Kanzi: Required for KANZI compression (best compression ratio)
  Download: https://github.com/flanglet/kanzi-java

Performance Notes:

• Streaming: Memory usage is bounded (~256KB) regardless of database size
• Parallel: Dump generation and SQL writing happen simultaneously
• Compression: Applied on-the-fly, no temporary uncompressed files
• Large Databases: Designed to handle multi-GB databases efficiently

Error Handling:

• Missing compression libraries fall back to uncompressed output
• Corrupted databases are processed in recovery mode
• Timeouts prevent indefinite hanging

Field Summary

Fields inherited from class org.h2.util.Tool

out

Constructor Summary

Constructors

Constructor	Description
DirectRecover()	Creates default instance

Method Summary

Modifier and Type	Method	Description
static void	execute(java.lang.String dir, java.lang.String db, java.io.PrintWriter writer)	Public method to process database files and write to a provided writer (pipe-like).
CompressionType[]	getAvailableCompressionTypes()	Get a list of available compression types based on classpath.
CompressionType	getCompressionType()	Gets the current compression type.
boolean	isCompressionAvailable(CompressionType type)	Check if a compression library is available without loading it.
static void	main(java.lang.String... args)	Options are case-sensitive.
void	runTool(java.lang.String... args)	Enhanced runTool method with compression support.
void	setCompressionType(CompressionType type)	Sets the compression type for SQL output files.

Methods inherited from class org.h2.tools.Recover

checkPowerOff, checkWritingAllowed, execute, getCompareMode, getDatabasePath, getLobFileListCache, getLobStorage, getLobSyncObject, getMaxLengthInplaceLob, getTempFileDeleter, openFile, readBlobMap, readClobMap, readLob

Methods inherited from class org.h2.util.Tool

getMainClassName, isOption, printNoDatabaseFilesFound, setOut, showUsage, showUsageAndThrowUnsupportedOption, throwUnsupportedOption

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DirectRecover

public DirectRecover()

Creates default instance

Method Detail

main

public static void main(java.lang.String... args)
                 throws java.sql.SQLException

Options are case-sensitive.

Supported options (in addition to base Recover options)

[-compress <type>]	Compress SQL output (none, gzip, zip, bzip2, kanzi, default: none)
[-trace]	Enable verbose debug output

Parameters:

args - the command line arguments

Throws:

java.sql.SQLException - on failure

runTool

public void runTool(java.lang.String... args)
             throws java.sql.SQLException

Enhanced runTool method with compression support.

Overrides:

runTool in class Recover

Parameters:

args - the command line arguments

Throws:

java.sql.SQLException - on failure

isCompressionAvailable

public boolean isCompressionAvailable(CompressionType type)

Check if a compression library is available without loading it.

Parameters:

type - the compression type to check

Returns:

true if the library is available, false otherwise

getAvailableCompressionTypes

public CompressionType[] getAvailableCompressionTypes()

Get a list of available compression types based on classpath.

Returns:

array of available compression types

setCompressionType

public void setCompressionType(CompressionType type)

Sets the compression type for SQL output files.

Parameters:

type - the compression type to use

getCompressionType

public CompressionType getCompressionType()

Gets the current compression type.

Returns:

the current compression type

execute

public static void execute(java.lang.String dir,
                           java.lang.String db,
                           java.io.PrintWriter writer)
                    throws java.sql.SQLException

Public method to process database files and write to a provided writer (pipe-like).

Parameters:

dir - the directory

db - the database name (null for all databases)

writer - the output writer

Throws:

java.sql.SQLException - on failure