unsorted/docsrc/guide/dbfiles.xml

<!-- $Id: dbfiles.xml,v 1.3 2005/04/14 21:52:34 unsaved Exp $ -->

<appendix>

<title>Hsqldb Database Files and Recovery</title>
<appendixinfo>
    <edition>$Revision: 1.3 $</edition>
    <pubdate>$Date: 2005/04/14 21:52:34 $</pubdate>
    <keywordset>
        <keyword>HSQLDB</keyword>
        <keyword>Data</keyword>
        <keyword>Files</keyword>
    </keywordset>
    <legalnotice><para>
        This text is based on HypersonicSQL documentation, updated to reflect 
        the latest version 1.8.0 of HSQLDB.
    </para></legalnotice>
</appendixinfo>

<section>
    <para>
        The Standalone and Client/Server modes will in most cases use files to 
        store all data to disk in a persistent and safe way.
        This document describes the meaning of the files, the states and the 
        procedures followed by the engine to recover the data.
    </para> <para>
        A database named 'test' is used in this description.
        The database files will be as follows.
    </para>
</section>

<variablelist>
    <title>Database Files</title>
    <varlistentry><term>test.properties</term><listitem><para>
        Contains the entry 'modified'.
        If the entry 'modified' is set to 'yes' then the database is either 
        running or was not closed correctly (because the close algorithm sets 
        'modified' to 'no' at the end).
        </para></listitem>
    </varlistentry>
    <varlistentry><term>test.script</term><listitem><para>
        This file contains the SQL statements that makes up the database up to 
        the last checkpoint - it is in synch with
        <filename>test.backup</filename>.
        </para></listitem>
    </varlistentry>
    <varlistentry><term>test.data</term><listitem><para>
        This file contains the (binary) data records for CACHED tables only.
        </para></listitem>
    </varlistentry>
    <varlistentry><term>test.backup</term><listitem><para>
        This is compressed file that contains the complete backup of the old 
        <filename>test.data</filename> file at the time of last checkpoint.
        </para></listitem>
    </varlistentry>
    <varlistentry><term>test.log</term><listitem><para>
        This file contains the extra SQL statements that have modified the 
        database since the last checkpoint (something like the 'Redo-log' or 
        'Transaction-log', but just text).
        </para><para>
        In the above list, a checkpoint results from both a CHECKPOINT command 
        and a SHUTDOWN command.
        </para></listitem>
    </varlistentry>
</variablelist>

<section>
    <title>States</title>
    <section>Database is closed correctly</section>
    <itemizedlist>
        <title>State after using the command <literal>SHUTDOWN</literal></title>
        <listitem><para>
            The <filename>test.data</filename> file is fully updated.
        </para></listitem> <listitem><para>
            The <filename>test.backup</filename> contains the compressed 
            <filename>test.data</filename> file.
        </para></listitem> <listitem><para>
            The  <filename>test.script</filename> contains the information in 
            the database, excluding data for CACHED and TEXT tables. 
        </para></listitem> <listitem><para>
            The <filename>test.properties</filename> contains the entry 
            'modified'  set to 'no'.
        </para></listitem> <listitem><para>
            There is no <filename>test.log</filename> file.
        </para></listitem>
    </itemizedlist>
    <section>Database is closed correctly with SHUTDOWN SCRIPT</section>
    <itemizedlist>
        <title>State after using the command <literal>SHUTDOWN SCRIPT</literal></title>
        <listitem><para>
            The <filename>test.data</filename> file does not exist; all CACHED 
            table data is in the <filename>test.script</filename> file
        </para></listitem> <listitem><para>
            The <filename>test.backup</filename> does not exist.
        </para></listitem> <listitem><para>
            The  <filename>test.script</filename> contains the information in 
            the database, including data for CACHED and TEXT tables.
        </para></listitem> <listitem><para>
            The <filename>test.properties</filename> contains the entry 
            'modified'  set to 'no'.
        </para></listitem> <listitem><para>
            There is no <filename>test.log</filename> file.
        </para></listitem>
    </itemizedlist>
    <section>Database is aborted</section>
    <para>
        This may happen by sudden power off, Ctrl+C in Windows, but may be 
        simulated using the command SHUTDOWN IMMEDIATELY.
    </para>
    <itemizedlist>
        <title>Aborted Database state</title>
        <listitem><para>
            The <filename>test.properties</filename> still containes 
            'modified=yes'.
        </para></listitem> <listitem><para>
            The <filename>test.script</filename> contains a snapshot of the 
            database at the last checkpoint and is OK.
        </para></listitem> <listitem><para>
            The <filename>test.data</filename> file may be corrupt because the 
            cache in memory was not written out completely.
        </para></listitem> <listitem><para>
            The <filename>test.backup</filename> file contains a snapshot of 
            <filename>test.data</filename> that corresponds to 
            <filename>test.script</filename>.
        </para></listitem> <listitem><para>
            The  <filename>test.log</filename> file contain all information to 
            re-do all changes since the snanapshot.
            As a result of abnormal termination, this file may be partially 
            corrupt.
        </para></listitem>
    </itemizedlist>
</section>

<section>
    <title>Procedures</title>
    <para>
        The database engine performs the following procedures internally in 
        different circumstances. 
    </para>
    <section><title>Clean Shutdown</title>
        <procedure>
            <title>Clean Hsqldb database shutdown</title>
            <step><para>
                The <filename>test.data</filename> file is written completely 
                (all the modified cached table rows are witten out) and closed.
            </para></step> <step><para>
                The <filename>test.backup.new</filename> is created (containing 
                the compressed <filename>test.data</filename> file)
            </para></step> <step><para>
                The file <filename>test.script.new</filename> is created using 
                the information in the database (and thus shrinks because no 
                UPDATE and DELETE statements; only INSERT).
            </para></step> <step><para>
                The entry 'modified' in the properties file is set to 
                'yes-new-files'
            </para></step> <step><para>
                The file <filename>test.script</filename> is deleted
            </para></step> <step><para>
                The file <filename>test.script.new</filename> is renamed to 
                <filename>test.script</filename>
            </para></step> <step><para>
                The file <filename>test.backup</filename> is deleted
            </para></step> <step><para>
                The file <filename>test.backup.new</filename> is renamed to 
                <filename>test.backup</filename>
            </para></step> <step><para>
                The entry 'modified' in the properties file is set to 'no' 
            </para></step> <step><para>
                The file <filename>test.log</filename> is deleted
            </para></step>
        </procedure>
    </section>
    <section><title>Startup</title>
        <procedure>
            <title>Database is opened</title>
            <step><para>
                Check if the database files are in use (by checking a special 
                <filename>test.lck</filename> file).
            </para></step> <step><para>
                See if the <filename>test.properties</filename> file exists,
                otherwise create it.
            </para></step> <step><para>
                If the <filename>test.properties</filename> did not exist, then 
                this is a new database.
                Create the empty <filename>test.log</filename> to append new 
                commands.
            </para></step> <step><para>
                If it is an existing database, check in the
                <filename>test.properties</filename> file if 'modified=yes'.
                This would mean last time it was not closed correctly, and thus 
                the <filename>test.data</filename> file may be corrupted or 
                incomplete.
                In this case the 'REPAIR' algorithm is executed (see below), 
                before the database is opened normally.
            </para></step> <step><para>
                Otherwise, if in the <filename>test.properties</filename> file
                'modified=yes-new-files', then the (old) 
                <filename>test.backup</filename> and 
                <filename>test.script</filename> files are deleted and the new 
                <filename>test.script.new</filename> file is renamed to 
                <filename>test.script</filename>.
                </para></step> <step><para>
                Open the <filename>test.script</filename> file and execute the 
                commands.
            </para></step> <step><para>
                Create the empty test.log to append new commands.
            </para></step>
        </procedure>
    </section>
    <section><title>Repair</title>
        <para>
            The current <filename>test.data</filename> file is corrupt, but 
            with the old <filename>test.data</filename> (from the 
            <filename>test.backup</filename> file and 
            <filename>test.script</filename>) and the current 
            <filename>test.log</filename>, the database is made up-to-date.
            The database engine takes these steps:
        </para>
        <procedure>
            <title>Database Repair</title>
            <step><para>
                    Restore the old <filename>test.data</filename> file from 
                    the backup (uncompress the <filename>test.backup</filename>
                    and overwrite <filename>test.data</filename>).
            </para></step><step><para>
                Execute all commands in the <filename>test.script</filename> 
                file.
            </para></step><step><para>
                Execute all commands in the <filename>test.log</filename> file.
                If due to corruption, an exception is thrown, the rest of the 
                lines of command in the <filename>test.log</filename> file are 
                ignored.
            </para></step><step><para>
                Close the database correctly (including a backup).
            </para></step>
        </procedure>
    </section>
</section>
</appendix>