Re: We really ought to do something about O_DIRECT and data=journalled on ext4

Поиск
Список
Период
Сортировка
От Greg Smith
Тема Re: We really ought to do something about O_DIRECT and data=journalled on ext4
Дата
Msg-id 4CFD782A.5020205@2ndquadrant.com
обсуждение исходный текст
Ответ на Re: We really ought to do something about O_DIRECT and data=journalled on ext4  (Tom Lane <tgl@sss.pgh.pa.us>)
Ответы Re: We really ought to do something about O_DIRECT and data=journalled on ext4  (Steve Singer <ssinger@ca.afilias.info>)
Re: We really ought to do something about O_DIRECT and data=journalled on ext4  (Tom Lane <tgl@sss.pgh.pa.us>)
Re: We really ought to do something about O_DIRECT and data=journalled on ext4  (Josh Berkus <josh@agliodbs.com>)
Список pgsql-hackers
Tom Lane wrote:<br /><blockquote cite="mid:1909.1291668822@sss.pgh.pa.us" type="cite"><pre wrap="">The various testing
that'sbeen reported so far is all for
 
Linux and thus doesn't directly address the question of whether other
kernels will have similar performance properties.</pre></blockquote><br /> Survey of some popular platforms:<br /><br
/>Linux:  don't want O_DIRECT by default for reliability reasons, and there's no clear performance win in the default
configwith small wal_buffers<br /><br /> Solaris:  O_DIRECT doesn't work, there's another API support has never been
addedfor; see <a class="moz-txt-link-freetext"
href="http://blogs.sun.com/jkshah/entry/postgresql_wal_sync_method_and">http://blogs.sun.com/jkshah/entry/postgresql_wal_sync_method_and</a><br
/><br/> Windows:  Small reported gains for O_DIRECT, i.e 10% at <a class="moz-txt-link-freetext"
href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg01615.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg01615.php</a><br
/><br/> FreeBSD:  It probably works there, but I've never seen good performance tests of it on this platform.<br /><br
/>Mac OS X:  Like Solaris, there's a similar mechanism but it's not O_DIRECT; see <a class="moz-txt-link-freetext"
href="http://stackoverflow.com/questions/2299402/how-does-one-do-raw-io-on-mac-os-x-ie-equivalent-to-linuxs-o-direct-flag">http://stackoverflow.com/questions/2299402/how-does-one-do-raw-io-on-mac-os-x-ie-equivalent-to-linuxs-o-direct-flag</a>
fornotes about the <span style="font-family: georgia;">F_NOCACHE  feature used.  Same basic situation as Solaris;
there'san API, but PostgreSQL doesn't use it yet.<br /><br /> So my guess is that some small percentage of Windows
usersmight notice a change here, and some testing on FreeBSD would be useful too.  That's about it for platforms that I
thinkanybody needs to worry about.<br /></span><br /><pre class="moz-signature" cols="72">-- 
 
Greg Smith   2ndQuadrant US    <a class="moz-txt-link-abbreviated"
href="mailto:greg@2ndQuadrant.com">greg@2ndQuadrant.com</a>  Baltimore, MD
 
PostgreSQL Training, Services and Support        <a class="moz-txt-link-abbreviated"
href="http://www.2ndQuadrant.us">www.2ndQuadrant.us</a>
"PostgreSQL 9.0 High Performance": <a class="moz-txt-link-freetext"
href="http://www.2ndQuadrant.com/books">http://www.2ndQuadrant.com/books</a>
</pre>

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Josh Berkus
Дата:
Сообщение: Re: WIP patch for parallel pg_dump
Следующее
От: flyusa2010 fly
Дата:
Сообщение: the number of file descriptors when using POSIX semaphore