@@ -25,7 +25,7 @@ ____________________________________________________________________________-->
2525<!--==========================orignal english content==========================
2626 <para>
2727 Reliability is an important property of any serious database
28- system, and <productname>PostgreSQL</> does everything possible to
28+ system, and <productname>PostgreSQL</productname > does everything possible to
2929 guarantee reliable operation. One aspect of reliable operation is
3030 that all data recorded by a committed transaction should be stored
3131 in a nonvolatile area that is safe from power loss, operating
@@ -51,8 +51,8 @@ ____________________________________________________________________________-->
5151 First, there is the operating system's buffer cache, which caches
5252 frequently requested disk blocks and combines disk writes. Fortunately,
5353 all operating systems give applications a way to force writes from
54- the buffer cache to disk, and <productname>PostgreSQL</> uses those
55- features. (See the <xref linkend="guc-wal-sync-method"> parameter
54+ the buffer cache to disk, and <productname>PostgreSQL</productname > uses those
55+ features. (See the <xref linkend="guc-wal-sync-method"/ > parameter
5656 to adjust how this is done.)
5757 </para>
5858____________________________________________________________________________-->
@@ -63,14 +63,14 @@ ____________________________________________________________________________-->
6363<!--==========================orignal english content==========================
6464 <para>
6565 Next, there might be a cache in the disk drive controller; this is
66- particularly common on <acronym>RAID</> controller cards. Some of
67- these caches are <firstterm>write-through</>, meaning writes are sent
66+ particularly common on <acronym>RAID</acronym > controller cards. Some of
67+ these caches are <firstterm>write-through</firstterm >, meaning writes are sent
6868 to the drive as soon as they arrive. Others are
69- <firstterm>write-back</>, meaning data is sent to the drive at
69+ <firstterm>write-back</firstterm >, meaning data is sent to the drive at
7070 some later time. Such caches can be a reliability hazard because the
7171 memory in the disk controller cache is volatile, and will lose its
7272 contents in a power failure. Better controller cards have
73- <firstterm>battery-backup units</> (<acronym>BBU</>s), meaning
73+ <firstterm>battery-backup units</
firstterm > (<acronym>BBU</
acronym >s), meaning
7474 the card has a battery that
7575 maintains power to the cache in case of system power loss. After power
7676 is restored the data will be written to the disk drives.
@@ -222,11 +222,11 @@ ____________________________________________________________________________-->
222222 If you use SSDs, be aware that many of these do not honor cache flush
223223 commands by default.
224224 You can test for reliable I/O subsystem behavior using <ulink
225- url="http ://brad.livejournal.com/2116715.html"><filename>diskchecker.pl</filename></ulink>.
225+ url="https ://brad.livejournal.com/2116715.html"><filename>diskchecker.pl</filename></ulink>.
226226 </para>
227227____________________________________________________________________________-->
228228 <para>
229- 在操作系统向存储硬件发出一个写请求的时候，它没有什么好办法来保证数据真正到达非易失的存储区域。 实际上，确保所有存储部件都保证数据和文件系统元数据的完整性是管理员的责任。 避免使用那些没有电池作为后备的写高速缓存的磁盘控制器。在驱动器级别，如果驱动器不能保证在关闭（掉电）之前写入数据， 那么关闭回写高速缓冲。如果你在使用SSD，注意很多SSD默认都没有兑现高速缓存刷写命令。你可以使用<ulink url="http ://brad.livejournal.com/2116715.html"><filename>diskchecker.pl</filename></ulink>来测试可靠的I/O子系统行为。
229+ 在操作系统向存储硬件发出一个写请求的时候，它没有什么好办法来保证数据真正到达非易失的存储区域。 实际上，确保所有存储部件都保证数据和文件系统元数据的完整性是管理员的责任。 避免使用那些没有电池作为后备的写高速缓存的磁盘控制器。在驱动器级别，如果驱动器不能保证在关闭（掉电）之前写入数据， 那么关闭回写高速缓冲。如果你在使用SSD，注意很多SSD默认都没有兑现高速缓存刷写命令。你可以使用<ulink url="https ://brad.livejournal.com/2116715.html"><filename>diskchecker.pl</filename></ulink>来测试可靠的I/O子系统行为。
230230 </para>
231231
232232<!--==========================orignal english content==========================
@@ -1054,7 +1054,7 @@ ____________________________________________________________________________-->
10541054 </para>
10551055____________________________________________________________________________-->
10561056 <para>
1057- <acronym>WAL</acronym>是自动被启用的。
除了做一些设置满足存放 <acronym>WAL</acronym>
日志的磁盘空间需求以及一些必要的调节以外 （参阅<xref linkend="wal-configuration"/>），
对管理员没有什么其他要求 。
1057+ <acronym>WAL</acronym>是自动被启用的。
除了确保满足 <acronym>WAL</acronym>
日志存放所需要的磁盘空间以及一些必要的调优外 （参阅<xref linkend="wal-configuration"/>），
管理员无需执行任何操作 。
10581058 </para>
10591059
10601060<!--==========================orignal english content==========================
@@ -1083,20 +1083,19 @@ ____________________________________________________________________________-->
10831083 <acronym>WAL</acronym> logs are stored in the directory
10841084 <filename>pg_wal</filename> under the data directory, as a set of
10851085 segment files, normally each 16 MB in size (but the size can be changed
1086- by altering the <option>-−with-wal-segsize</> configure option when
1087- building the server). Each segment is divided into pages, normally
1088- 8 kB each (this size can be changed via the <option>-−with-wal-blocksize</>
1089- configure option). The log record headers are described in
1090- <filename>access/xlogrecord.h</filename>; the record content is dependent
1091- on the type of event that is being logged. Segment files are given
1092- ever-increasing numbers as names, starting at
1086+ by altering the <option>-−with-wal-segsize</> initdb option). Each segment is
1087+ divided into pages, normally 8 kB each (this size can be changed via the
1088+ <option>-−with-wal-blocksize</> configure option). The log record headers
1089+ are described in <filename>access/xlogrecord.h</filename>; the record
1090+ content is dependent on the type of event that is being logged. Segment
1091+ files are given ever-increasing numbers as names, starting at
10931092 <filename>000000010000000000000000</filename>. The numbers do not wrap,
10941093 but it will take a very, very long time to exhaust the
10951094 available stock of numbers.
10961095 </para>
10971096____________________________________________________________________________-->
10981097 <para>
1099- <acronym>WAL</acronym>日志被存放在数据目录的<filename>pg_wal</filename>目录里，它是作为一个文件段的集合存储的，通常每个段16MB大（但是可以在构建服务器时通过修改 <option>--with-wal-segsize</option>配置选项来修改段的大小 ）。每个段分割成多个页，通常每个页为8K（该尺寸可以通过<option>--with-wal-blocksize</option>配置选项来修改）。日志记录头部在<filename>access/xlogrecord.h</filename>里描述；日志内容取决于它记录的事件类型。段文件的名字是不断增长的数字，从<filename>000000010000000000000000</filename>开始。目前这些数字不能回卷，不过要把所有可用的数字都用光也需要非常非常长的时间。
1098+ <acronym>WAL</acronym>日志被存放在数据目录的<filename>pg_wal</filename>目录里，它是作为一个文件段的集合存储的，通常每个段16MB大小（不过这个大小可以通过initdb配置选项 <option>--with-wal-segsize</option>来修改 ）。每个段分割成多个页，通常每个页为8K（该尺寸可以通过<option>--with-wal-blocksize</option>配置选项来修改）。日志记录头部在<filename>access/xlogrecord.h</filename>里描述；日志内容取决于它记录的事件类型。段文件的名字是不断增长的数字，从<filename>000000010000000000000000</filename>开始。目前这些数字不能回卷，不过要把所有可用的数字都用光也需要非常非常长的时间。
11001099 </para>
11011100
11021101<!--==========================orignal english content==========================
0 commit comments