Improvements to the backup & restore documentation.

doc/src/sgml/backup.sgml

@@ -1,5 +1,5 @@
 <!--
-$PostgreSQL: pgsql/doc/src/sgml/backup.sgml,v 2.38 2004/03/09 16:57:46 neilc Exp $
+$PostgreSQL: pgsql/doc/src/sgml/backup.sgml,v 2.39 2004/04/22 07:02:35 neilc Exp $
 -->
 <chapter id="backup">
  <title>Backup and Restore</title>
@@ -30,7 +30,7 @@ $PostgreSQL: pgsql/doc/src/sgml/backup.sgml,v 2.38 2004/03/09 16:57:46 neilc Exp
    commands that, when fed back to the server, will recreate the
    database in the same state as it was at the time of the dump.
    <productname>PostgreSQL</> provides the utility program
-   <application>pg_dump</> for this purpose. The basic usage of this
+   <xref linkend="app-pgdump"> for this purpose. The basic usage of this
    command is:
 <synopsis>
 pg_dump <replaceable class="parameter">dbname</replaceable> &gt; <replaceable class="parameter">outfile</replaceable>
@@ -126,10 +126,11 @@ psql <replaceable class="parameter">dbname</replaceable> &lt; <replaceable class
    </para>
 
    <para>
-    Once restored, it is wise to run <command>ANALYZE</> on each
+    Once restored, it is wise to run <xref linkend="sql-analyze"
-    database so the optimizer has useful statistics. You
+    endterm="sql-analyze-title"> on each database so the optimizer has
-    can also run <command>vacuumdb -a -z</> to <command>ANALYZE</> all
+    useful statistics. You can also run <command>vacuumdb -a -z</> to
-    databases.
+    <command>VACUUM ANALYZE</> all databases; this is equivalent to
+    running <command>VACUUM ANALYZE</command> manually.
    </para>
 
    <para>
@@ -153,13 +154,11 @@ pg_dump -h <replaceable>host1</> <replaceable>dbname</> | psql -h <replaceable>h
     </para>
    </important>
 
-   <tip>
    <para>
-    Restore performance can be improved by increasing the
+    For advice on how to load large amounts of data into
-    configuration parameter <xref
+    <productname>PostgreSQL</productname> efficiently, refer to <xref
-    linkend="guc-maintenance-work-mem">.
+    linkend="populate">.
    </para>
-   </tip>
   </sect2>
 
   <sect2 id="backup-dump-all">
@@ -167,12 +166,11 @@ pg_dump -h <replaceable>host1</> <replaceable>dbname</> | psql -h <replaceable>h
 
    <para>
     The above mechanism is cumbersome and inappropriate when backing
-    up an entire database cluster. For this reason the
+    up an entire database cluster. For this reason the <xref
-    <application>pg_dumpall</> program is provided.
+    linkend="app-pg-dumpall"> program is provided.
     <application>pg_dumpall</> backs up each database in a given
-    cluster, and also preserves cluster-wide data such as
+    cluster, and also preserves cluster-wide data such as users and
-    users and groups. The call sequence for
+    groups. The basic usage of this command is:
-    <application>pg_dumpall</> is simply
 <synopsis>
 pg_dumpall &gt; <replaceable>outfile</>
 </synopsis>
@@ -195,7 +193,7 @@ psql template1 &lt; <replaceable class="parameter">infile</replaceable>
     Since <productname>PostgreSQL</productname> allows tables larger
     than the maximum file size on your system, it can be problematic
     to dump such a table to a file, since the resulting file will likely
-    be larger than the maximum size allowed by your system. As
+    be larger than the maximum size allowed by your system. Since
     <application>pg_dump</> can write to the standard output, you can
     just use standard Unix tools to work around this possible problem.
    </para>
@@ -274,7 +272,7 @@ pg_dump -Fc <replaceable class="parameter">dbname</replaceable> > <replaceable c
     For reasons of backward compatibility, <application>pg_dump</>
     does not dump large objects by default.<indexterm><primary>large
     object</primary><secondary>backup</secondary></indexterm> To dump
-    large objects you must use either the custom or the TAR output
+    large objects you must use either the custom or the tar output
     format, and use the <option>-b</> option in
     <application>pg_dump</>. See the reference pages for details.  The
     directory <filename>contrib/pg_dumplo</> of the
@@ -315,11 +313,12 @@ tar -cf backup.tar /usr/local/pgsql/data
      <para>
       The database server <emphasis>must</> be shut down in order to
       get a usable backup. Half-way measures such as disallowing all
-      connections will not work as there is always some buffering
+      connections will <emphasis>not</emphasis> work
-      going on. Information about stopping the server can be
+      (<command>tar</command> and similar tools do not take an atomic
-      found in <xref linkend="postmaster-shutdown">.  Needless to say
+      snapshot of the state of the filesystem at a point in
-      that you also need to shut down the server before restoring the
+      time). Information about stopping the server can be found in
-      data.
+      <xref linkend="postmaster-shutdown">.  Needless to say that you
+      also need to shut down the server before restoring the data.
      </para>
     </listitem>
 

doc/src/sgml/perform.sgml

@@ -1,5 +1,5 @@
 <!--
-$PostgreSQL: pgsql/doc/src/sgml/perform.sgml,v 1.43 2004/03/25 18:57:57 tgl Exp $
+$PostgreSQL: pgsql/doc/src/sgml/perform.sgml,v 1.44 2004/04/22 07:02:36 neilc Exp $
 -->
 
  <chapter id="performance-tips">
@@ -28,8 +28,8 @@ $PostgreSQL: pgsql/doc/src/sgml/perform.sgml,v 1.43 2004/03/25 18:57:57 tgl Exp
     plan</firstterm> for each query it is given.  Choosing the right
     plan to match the query structure and the properties of the data
     is absolutely critical for good performance.  You can use the
-    <command>EXPLAIN</command> command to see what query plan the system
+    <xref linkend="sql-explain" endterm="sql-explain-title"> command
-    creates for any query.
+    to see what query plan the system creates for any query.
     Plan-reading is an art that deserves an extensive tutorial, which
     this is not; but here is some basic information.
    </para>
@@ -638,30 +638,51 @@ SELECT * FROM x, y, a, b, c WHERE something AND somethingelse;
    </indexterm>
 
    <para>
-    Turn off autocommit and just do one commit at
+    Turn off autocommit and just do one commit at the end.  (In plain
-    the end.  (In plain SQL, this means issuing <command>BEGIN</command>
+    SQL, this means issuing <command>BEGIN</command> at the start and
-    at the start and <command>COMMIT</command> at the end.  Some client
+    <command>COMMIT</command> at the end.  Some client libraries may
-    libraries may do this behind your back, in which case you need to
+    do this behind your back, in which case you need to make sure the
-    make sure the library does it when you want it done.)
+    library does it when you want it done.)  If you allow each
-    If you allow each insertion to be committed separately,
+    insertion to be committed separately,
-    <productname>PostgreSQL</productname> is doing a lot of work for each
+    <productname>PostgreSQL</productname> is doing a lot of work for
-    row that is added.
+    each row that is added.  An additional benefit of doing all
-    An additional benefit of doing all insertions in one transaction
+    insertions in one transaction is that if the insertion of one row
-    is that if the insertion of one row were to fail then the
+    were to fail then the insertion of all rows inserted up to that
-    insertion of all rows inserted up to that point would be rolled
+    point would be rolled back, so you won't be stuck with partially
-    back, so you won't be stuck with partially loaded data.
+    loaded data.
+   </para>
+
+   <para>
+    If you are issuing a large sequence of <command>INSERT</command>
+    commands to bulk load some data, also consider using <xref
+    linkend="sql-prepare" endterm="sql-prepare-title"> to create a
+    prepared <command>INSERT</command> statement. Since you are
+    executing the same command multiple times, it is more efficient to
+    prepare the command once and then use <command>EXECUTE</command>
+    as many times as required.
    </para>
   </sect2>
 
   <sect2 id="populate-copy-from">
-   <title>Use <command>COPY FROM</command></title>
+   <title>Use <command>COPY</command></title>
 
    <para>
-    Use <command>COPY FROM STDIN</command> to load all the rows in one
+    Use <xref linkend="sql-copy" endterm="sql-copy-title"> to load
-    command, instead of using a series of <command>INSERT</command>
+    all the rows in one command, instead of using a series of
-    commands.  This reduces parsing, planning, etc.  overhead a great
+    <command>INSERT</command> commands.  The <command>COPY</command>
-    deal. If you do this then it is not necessary to turn off
+    command is optimized for loading large numbers of rows; it is less
-    autocommit, since it is only one command anyway.
+    flexible than <command>INSERT</command>, but incurs significantly
+    less overhead for large data loads. Since <command>COPY</command>
+    is a single command, there is no need to disable autocommit if you
+    use this method to populate a table.
+   </para>
+
+   <para>
+    Note that loading a large number of rows using
+    <command>COPY</command> is almost always faster than using
+    <command>INSERT</command>, even if multiple
+    <command>INSERT</command> commands are batched into a single
+    transaction.
    </para>
   </sect2>
 
@@ -678,11 +699,12 @@ SELECT * FROM x, y, a, b, c WHERE something AND somethingelse;
 
    <para>
     If you are augmenting an existing table, you can drop the index,
-    load the table, then recreate the index. Of
+    load the table, and then recreate the index. Of course, the
-    course, the database performance for other users may be adversely 
+    database performance for other users may be adversely affected
-    affected during the time that the index is missing.  One should also
+    during the time that the index is missing.  One should also think
-    think twice before dropping unique indexes, since the error checking
+    twice before dropping unique indexes, since the error checking
-    afforded by the unique constraint will be lost while the index is missing.
+    afforded by the unique constraint will be lost while the index is
+    missing.
    </para>
   </sect2>
 
@@ -701,16 +723,39 @@ SELECT * FROM x, y, a, b, c WHERE something AND somethingelse;
    </para>
   </sect2>
 
+  <sect2 id="populate-checkpoint-segments">
+   <title>Increase <varname>checkpoint_segments</varname></title>
+
+   <para>
+    Temporarily increasing the <xref
+    linkend="guc-checkpoint-segments"> configuration variable can also
+    make large data loads faster.  This is because loading a large
+    amount of data into <productname>PostgreSQL</productname> can
+    cause checkpoints to occur more often than the normal checkpoint
+    frequency (specified by the <varname>checkpoint_timeout</varname>
+    configuration variable). Whenever a checkpoint occurs, all dirty
+    pages must be flushed to disk. By increasing
+    <varname>checkpoint_segments</varname> temporarily during bulk
+    data loads, the number of checkpoints that are required can be
+    reduced.
+   </para>
+  </sect2>
+
   <sect2 id="populate-analyze">
    <title>Run <command>ANALYZE</command> Afterwards</title>
 
    <para>
-    It's a good idea to run <command>ANALYZE</command> or <command>VACUUM
+    Whenever you have significantly altered the distribution of data
-    ANALYZE</command> anytime you've added or updated a lot of data,
+    within a table, running <xref linkend="sql-analyze"
-    including just after initially populating a table.  This ensures that
+    endterm="sql-analyze-title"> is strongly recommended. This
-    the planner has up-to-date statistics about the table.  With no statistics
+    includes when bulk loading large amounts of data into
-    or obsolete statistics, the planner may make poor choices of query plans,
+    <productname>PostgreSQL</productname>.  Running
-    leading to bad performance on queries that use your table.
+    <command>ANALYZE</command> (or <command>VACUUM ANALYZE</command>)
+    ensures that the planner has up-to-date statistics about the
+    table.  With no statistics or obsolete statistics, the planner may
+    make poor decisions during query planning, leading to poor
+    performance on any tables with inaccurate or nonexistent
+    statistics.
    </para>
   </sect2>
   </sect1>