1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
|
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>27.1. Comparison of Different Solutions</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="high-availability.html" title="Chapter 27. High Availability, Load Balancing, and Replication" /><link rel="next" href="warm-standby.html" title="27.2. Log-Shipping Standby Servers" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">27.1. Comparison of Different Solutions</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="high-availability.html" title="Chapter 27. High Availability, Load Balancing, and Replication">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="high-availability.html" title="Chapter 27. High Availability, Load Balancing, and Replication">Up</a></td><th width="60%" align="center">Chapter 27. High Availability, Load Balancing, and Replication</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="warm-standby.html" title="27.2. Log-Shipping Standby Servers">Next</a></td></tr></table><hr /></div><div class="sect1" id="DIFFERENT-REPLICATION-SOLUTIONS"><div class="titlepage"><div><div><h2 class="title" style="clear: both">27.1. Comparison of Different Solutions</h2></div></div></div><div class="variablelist"><dl class="variablelist"><dt><span class="term">Shared Disk Failover</span></dt><dd><p>
Shared disk failover avoids synchronization overhead by having only one
copy of the database. It uses a single disk array that is shared by
multiple servers. If the main database server fails, the standby server
is able to mount and start the database as though it were recovering from
a database crash. This allows rapid failover with no data loss.
</p><p>
Shared hardware functionality is common in network storage devices.
Using a network file system is also possible, though care must be
taken that the file system has full <acronym class="acronym">POSIX</acronym> behavior (see <a class="xref" href="creating-cluster.html#CREATING-CLUSTER-NFS" title="19.2.2.1. NFS">Section 19.2.2.1</a>). One significant limitation of this
method is that if the shared disk array fails or becomes corrupt, the
primary and standby servers are both nonfunctional. Another issue is
that the standby server should never access the shared storage while
the primary server is running.
</p></dd><dt><span class="term">File System (Block Device) Replication</span></dt><dd><p>
A modified version of shared hardware functionality is file system
replication, where all changes to a file system are mirrored to a file
system residing on another computer. The only restriction is that
the mirroring must be done in a way that ensures the standby server
has a consistent copy of the file system — specifically, writes
to the standby must be done in the same order as those on the primary.
<span class="productname">DRBD</span> is a popular file system replication solution
for Linux.
</p></dd><dt><span class="term">Write-Ahead Log Shipping</span></dt><dd><p>
Warm and hot standby servers can be kept current by reading a
stream of write-ahead log (<acronym class="acronym">WAL</acronym>)
records. If the main server fails, the standby contains
almost all of the data of the main server, and can be quickly
made the new primary database server. This can be synchronous or
asynchronous and can only be done for the entire database server.
</p><p>
A standby server can be implemented using file-based log shipping
(<a class="xref" href="warm-standby.html" title="27.2. Log-Shipping Standby Servers">Section 27.2</a>) or streaming replication (see
<a class="xref" href="warm-standby.html#STREAMING-REPLICATION" title="27.2.5. Streaming Replication">Section 27.2.5</a>), or a combination of both. For
information on hot standby, see <a class="xref" href="hot-standby.html" title="27.4. Hot Standby">Section 27.4</a>.
</p></dd><dt><span class="term">Logical Replication</span></dt><dd><p>
Logical replication allows a database server to send a stream of data
modifications to another server. <span class="productname">PostgreSQL</span>
logical replication constructs a stream of logical data modifications
from the WAL. Logical replication allows replication of data changes on
a per-table basis. In addition, a server that is publishing its own
changes can also subscribe to changes from another server, allowing data
to flow in multiple directions. For more information on logical
replication, see <a class="xref" href="logical-replication.html" title="Chapter 31. Logical Replication">Chapter 31</a>. Through the
logical decoding interface (<a class="xref" href="logicaldecoding.html" title="Chapter 49. Logical Decoding">Chapter 49</a>),
third-party extensions can also provide similar functionality.
</p></dd><dt><span class="term">Trigger-Based Primary-Standby Replication</span></dt><dd><p>
A trigger-based replication setup typically funnels data modification
queries to a designated primary server. Operating on a per-table basis,
the primary server sends data changes (typically) asynchronously to the
standby servers. Standby servers can answer queries while the primary is
running, and may allow some local data changes or write activity. This
form of replication is often used for offloading large analytical or data
warehouse queries.
</p><p>
<span class="productname">Slony-I</span> is an example of this type of
replication, with per-table granularity, and support for multiple standby
servers. Because it updates the standby server asynchronously (in
batches), there is possible data loss during fail over.
</p></dd><dt><span class="term">SQL-Based Replication Middleware</span></dt><dd><p>
With SQL-based replication middleware, a program intercepts
every SQL query and sends it to one or all servers. Each server
operates independently. Read-write queries must be sent to all servers,
so that every server receives any changes. But read-only queries can be
sent to just one server, allowing the read workload to be distributed
among them.
</p><p>
If queries are simply broadcast unmodified, functions like
<code class="function">random()</code>, <code class="function">CURRENT_TIMESTAMP</code>, and
sequences can have different values on different servers.
This is because each server operates independently, and because
SQL queries are broadcast rather than actual data changes. If
this is unacceptable, either the middleware or the application
must determine such values from a single source and then use those
values in write queries. Care must also be taken that all
transactions either commit or abort on all servers, perhaps
using two-phase commit (<a class="xref" href="sql-prepare-transaction.html" title="PREPARE TRANSACTION"><span class="refentrytitle">PREPARE TRANSACTION</span></a>
and <a class="xref" href="sql-commit-prepared.html" title="COMMIT PREPARED"><span class="refentrytitle">COMMIT PREPARED</span></a>).
<span class="productname">Pgpool-II</span> and <span class="productname">Continuent Tungsten</span>
are examples of this type of replication.
</p></dd><dt><span class="term">Asynchronous Multimaster Replication</span></dt><dd><p>
For servers that are not regularly connected or have slow
communication links, like laptops or
remote servers, keeping data consistent among servers is a
challenge. Using asynchronous multimaster replication, each
server works independently, and periodically communicates with
the other servers to identify conflicting transactions. The
conflicts can be resolved by users or conflict resolution rules.
Bucardo is an example of this type of replication.
</p></dd><dt><span class="term">Synchronous Multimaster Replication</span></dt><dd><p>
In synchronous multimaster replication, each server can accept
write requests, and modified data is transmitted from the
original server to every other server before each transaction
commits. Heavy write activity can cause excessive locking and
commit delays, leading to poor performance. Read requests can
be sent to any server. Some implementations use shared disk
to reduce the communication overhead. Synchronous multimaster
replication is best for mostly read workloads, though its big
advantage is that any server can accept write requests —
there is no need to partition workloads between primary and
standby servers, and because the data changes are sent from one
server to another, there is no problem with non-deterministic
functions like <code class="function">random()</code>.
</p><p>
<span class="productname">PostgreSQL</span> does not offer this type of replication,
though <span class="productname">PostgreSQL</span> two-phase commit (<a class="xref" href="sql-prepare-transaction.html" title="PREPARE TRANSACTION"><span class="refentrytitle">PREPARE TRANSACTION</span></a> and <a class="xref" href="sql-commit-prepared.html" title="COMMIT PREPARED"><span class="refentrytitle">COMMIT PREPARED</span></a>)
can be used to implement this in application code or middleware.
</p></dd></dl></div><p>
<a class="xref" href="different-replication-solutions.html#HIGH-AVAILABILITY-MATRIX" title="Table 27.1. High Availability, Load Balancing, and Replication Feature Matrix">Table 27.1</a> summarizes
the capabilities of the various solutions listed above.
</p><div class="table" id="HIGH-AVAILABILITY-MATRIX"><p class="title"><strong>Table 27.1. High Availability, Load Balancing, and Replication Feature Matrix</strong></p><div class="table-contents"><table class="table" summary="High Availability, Load Balancing, and Replication Feature Matrix" border="1"><colgroup><col class="col1" /><col class="col2" /><col class="col3" /><col class="col4" /><col class="col5" /><col class="col6" /><col class="col7" /><col class="col8" /><col class="col9" /></colgroup><thead><tr><th>Feature</th><th>Shared Disk</th><th>File System Repl.</th><th>Write-Ahead Log Shipping</th><th>Logical Repl.</th><th>Trigger-Based Repl.</th><th>SQL Repl. Middle-ware</th><th>Async. MM Repl.</th><th>Sync. MM Repl.</th></tr></thead><tbody><tr><td>Popular examples</td><td align="center">NAS</td><td align="center">DRBD</td><td align="center">built-in streaming repl.</td><td align="center">built-in logical repl., pglogical</td><td align="center">Londiste, Slony</td><td align="center">pgpool-II</td><td align="center">Bucardo</td><td align="center"> </td></tr><tr><td>Comm. method</td><td align="center">shared disk</td><td align="center">disk blocks</td><td align="center">WAL</td><td align="center">logical decoding</td><td align="center">table rows</td><td align="center">SQL</td><td align="center">table rows</td><td align="center">table rows and row locks</td></tr><tr><td>No special hardware required</td><td align="center"> </td><td align="center">•</td><td align="center">•</td><td align="center">•</td><td align="center">•</td><td align="center">•</td><td align="center">•</td><td align="center">•</td></tr><tr><td>Allows multiple primary servers</td><td align="center"> </td><td align="center"> </td><td align="center"> </td><td align="center">•</td><td align="center"> </td><td align="center">•</td><td align="center">•</td><td align="center">•</td></tr><tr><td>No overhead on primary</td><td align="center">•</td><td align="center"> </td><td align="center">•</td><td align="center">•</td><td align="center"> </td><td align="center">•</td><td align="center"> </td><td align="center"> </td></tr><tr><td>No waiting for multiple servers</td><td align="center">•</td><td align="center"> </td><td align="center">with sync off</td><td align="center">with sync off</td><td align="center">•</td><td align="center"> </td><td align="center">•</td><td align="center"> </td></tr><tr><td>Primary failure will never lose data</td><td align="center">•</td><td align="center">•</td><td align="center">with sync on</td><td align="center">with sync on</td><td align="center"> </td><td align="center">•</td><td align="center"> </td><td align="center">•</td></tr><tr><td>Replicas accept read-only queries</td><td align="center"> </td><td align="center"> </td><td align="center">with hot standby</td><td align="center">•</td><td align="center">•</td><td align="center">•</td><td align="center">•</td><td align="center">•</td></tr><tr><td>Per-table granularity</td><td align="center"> </td><td align="center"> </td><td align="center"> </td><td align="center">•</td><td align="center">•</td><td align="center"> </td><td align="center">•</td><td align="center">•</td></tr><tr><td>No conflict resolution necessary</td><td align="center">•</td><td align="center">•</td><td align="center">•</td><td align="center"> </td><td align="center">•</td><td align="center">•</td><td align="center"> </td><td align="center">•</td></tr></tbody></table></div></div><br class="table-break" /><p>
There are a few solutions that do not fit into the above categories:
</p><div class="variablelist"><dl class="variablelist"><dt><span class="term">Data Partitioning</span></dt><dd><p>
Data partitioning splits tables into data sets. Each set can
be modified by only one server. For example, data can be
partitioned by offices, e.g., London and Paris, with a server
in each office. If queries combining London and Paris data
are necessary, an application can query both servers, or
primary/standby replication can be used to keep a read-only copy
of the other office's data on each server.
</p></dd><dt><span class="term">Multiple-Server Parallel Query Execution</span></dt><dd><p>
Many of the above solutions allow multiple servers to handle multiple
queries, but none allow a single query to use multiple servers to
complete faster. This solution allows multiple servers to work
concurrently on a single query. It is usually accomplished by
splitting the data among servers and having each server execute its
part of the query and return results to a central server where they
are combined and returned to the user. This can be implemented using the
<span class="productname">PL/Proxy</span> tool set.
</p></dd></dl></div><p>
It should also be noted that because <span class="productname">PostgreSQL</span>
is open source and easily extended, a number of companies have
taken <span class="productname">PostgreSQL</span> and created commercial
closed-source solutions with unique failover, replication, and load
balancing capabilities. These are not discussed here.
</p></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="high-availability.html" title="Chapter 27. High Availability, Load Balancing, and Replication">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="high-availability.html" title="Chapter 27. High Availability, Load Balancing, and Replication">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="warm-standby.html" title="27.2. Log-Shipping Standby Servers">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Chapter 27. High Availability, Load Balancing, and Replication </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="40%" align="right" valign="top"> 27.2. Log-Shipping Standby Servers</td></tr></table></div></body></html>
|
