pgpool2.git / summary / log / commit / refs
commit f882eafda658b20cb00514b53ed06a38e0566709
Author: Tatsuo Ishii <ishii@sraoss.co.jp>
Date: Thu Apr 13 00:25:09 2023 +0000
Doc: add "Backing up PostgreSQL database" section to "Server Setup and Operation" chapter.
doc.ja/src/sgml/runtime.sgml | 57 ++++++++++++++++++++++++++++++++++++++++++++
doc/src/sgml/runtime.sgml | 38 +++++++++++++++++++++++++++++
2 files changed, 95 insertions(+)
diff --git a/doc.ja/src/sgml/runtime.sgml b/doc.ja/src/sgml/runtime.sgml
index 230320cbe..1e4b7e051 100644
--- a/doc.ja/src/sgml/runtime.sgml
+++ b/doc.ja/src/sgml/runtime.sgml
@@ -884,4 +884,61 @@ backend_flag0 = ALLOW_TO_FAILOVER
</sect2>
</sect1>
+ <sect1 id="backup">
+<!--
+ <title>Backing up PostgreSQL database</title>
+-->
+ <title>PostgreSQLデータベースのバックアップ</title>
+
+ <indexterm zone="backup">
+ <primary>backup</primary>
+ </indexterm>
+
+ <para>
+<!--
+ If you plan to backup <productname>PostgreSQL</productname>
+ database
+ using <command>pg_dump</command>, <command>pg_basebackup</command>
+ or any other tools, we strongly recommend to run the commands
+ against <productname>PostgreSQL</productname>
+ directly. As <productname>Pgpool-II</productname> is a proxy
+ software, it gives an overhead for relaying message packets. Since
+ obtaining a backup tends to produce lots of packets, performing
+ backup through <productname>Pgpool-II</productname> will be slow
+ compared with directly
+ connecting <productname>PostgreSQL</productname> unless the
+ database is very small.
+-->
+<command>pg_dump</command>、<command>pg_basebackup</command>その他のツールを使って<productname>PostgreSQL</productname>データベースをバックアップする際は、そうしたコマンドを<productname>PostgreSQL</productname>で直接実行することを強くお勧めします。
+<productname>Pgpool-II</productname>はproxyソフトウェアなので、メッセージパケットを転送する際にオーバヘッドが生じるからです。
+バックアップの取得は大量のパケットを生じるので、データベースが非常に小さくない限り、<productname>Pgpool-II</productname>経由でバックアップを取得するのは、直接<productname>PostgreSQL</productname>に接続するよりも遅くなります。
+ </para>
+ <para>
+<!--
+ Moreover, parallel <command>pg_dump</command> raises error if it is
+ executed through <productname>Pgpool-II</productname> because the
+ command handles snapshot id, which is a database depending object.
+-->
+ </para>
+ <para>
+ また、パラレル<command>pg_dump</command>では、データベース依存のオブジェクトであるスナップショットIDを扱うため、<productname>Pgpool-II</productname>経由で実行するとエラーになってしまいます。
+ </para>
+ <para>
+<!--
+ In most cases you want to choose primary server as the backup
+ target. If you want backup standby server, you have to be very
+ careful in choosing the right <productname>PostgreSQL</productname>
+ server to obtain a backup because if the data is outdated, you are
+ likely to have outdated database backup. You can
+ use <xref linkend="SQL-SHOW-POOL-NODES">
+ or <xref linkend="PCP-NODE-INFO"> to know how the standby server
+ catches up the primary server.
+-->
+ 多くの場合、プライマリサーバをバックアップ取得対象に選ぶと思います。
+ スタンバイサーバを選ぶ場合は、正しい<productname>PostgreSQL</productname>サーバを選ぶようにしましょう。
+ データが古ければ、バックアップも古いものになってしまう可能性が高いからです。
+ <xref linkend="SQL-SHOW-POOL-NODES">あるいは<xref linkend="PCP-NODE-INFO">を使って対象のスタンバイサーバがプライマリサーバに追いついているかどうかを知ることができます。
+ </para>
+ </sect1>
+
</chapter>
diff --git a/doc/src/sgml/runtime.sgml b/doc/src/sgml/runtime.sgml
index 34e0412a1..e923ce382 100644
--- a/doc/src/sgml/runtime.sgml
+++ b/doc/src/sgml/runtime.sgml
@@ -700,4 +700,42 @@ backend_flag0 = ALLOW_TO_FAILOVER
</sect1>
+ <sect1 id="backup">
+ <title>Backing up PostgreSQL database</title>
+
+ <indexterm zone="backup">
+ <primary>backup</primary>
+ </indexterm>
+
+ <para>
+ If you plan to backup <productname>PostgreSQL</productname>
+ database
+ using <command>pg_dump</command>, <command>pg_basebackup</command>
+ or any other tools, we strongly recommend to run the commands
+ against <productname>PostgreSQL</productname>
+ directly. As <productname>Pgpool-II</productname> is a proxy
+ software, it gives an overhead for relaying message packets. Since
+ obtaining a backup tends to produce lots of packets, performing
+ backup through <productname>Pgpool-II</productname> will be slow
+ compared with directly
+ connecting <productname>PostgreSQL</productname> unless the
+ database is very small.
+ </para>
+ <para>
+ Moreover, parallel <command>pg_dump</command> raises error if it is
+ executed through <productname>Pgpool-II</productname> because the
+ command handles snapshot id, which is a database depending object.
+ </para>
+ <para>
+ In most cases you want to choose primary server as the backup
+ target. If you want backup standby server, you have to be very
+ careful in choosing the right <productname>PostgreSQL</productname>
+ server to obtain a backup because if the data is outdated, you are
+ likely to have outdated database backup. You can
+ use <xref linkend="SQL-SHOW-POOL-NODES">
+ or <xref linkend="PCP-NODE-INFO"> to know how the standby server
+ catches up the primary server.
+ </para>
+ </sect1>
+
</chapter>
[parent: 3bd5f352d1da]