2009/10/20 James Cameron <span dir="ltr"><<a href="mailto:quozl@us.netrek.org">quozl@us.netrek.org</a>></span><br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On Tue, Oct 20, 2009 at 09:39:50AM +0800, Ray Rashif wrote:<br>
> Since on a Linux system files are written to almost every minute,<br>
> recording atime everytime an access occurs increases the disk I/O.<br>
<br>
</div>I'm not sure that is right; the way you put it.<br>
<br>
Perhaps you mean "Since on a Linux system files are read from almost<br>
every minute, recording atime every time a read occurs increases the<br>
write disk I/O."<br>
<br>
atime on a file is changed whenever the file is read. The change is<br>
buffered in memory for a while; usually no more than a minute, then is<br>
written out to disk.<br>
<br>
So atime causes at least one disk write I/O for a set of disk read I/Os.<br>
<br>
It certainly does impact performance ... if it is not needed.<br>
<div class="im"><br>
> I need to know access times for some files so I don't quite like that.<br>
> I use relatime: <a href="http://kerneltrap.org/node/14148" target="_blank">http://kerneltrap.org/node/14148</a><br>
<br>
</div>Another method is to place these files in a separate filesystem with<br>
atime enabled.</blockquote><div><br></div><div>I actually meant to type "read from", but thanks for correcting :) </div></div>