<div style="line-height:1.7;color:#000000;font-size:14px;font-family:Arial">Thank you for your reply.My glusterfs version is 3.3.2, the network.remote-dio and strict-o-direct didn't exist. In the mailing list you shared to me just now ,he was also confused that <span style="font-family: arial; line-height: normal; white-space: pre-wrap;">what is a surefire way to get direct-io-like behavior on gluster volume(s)? I mounted the glusterfs to the mountpoint(glusterfs) with direct-io-mode=enable,but when I executed "cp ./test1.mp4 /mnt/glusterfs/ " ,and then I executed "free -m " to check the size of cached, it got larger.The test1.mp4 is 2.5G, finally.the cached size is 2600M </span><br><br><br><br><br><div style="position:relative;zoom:1"></div><div id="divNeteaseMailCard"></div><br><pre><br>At 2016-03-22 16:20:44, "Raghavendra Gowdappa" <rgowdapp@redhat.com> wrote:
>
>
>----- Original Message -----
>> From: "Keiviw" <keiviw@163.com>
>> To: gluster-devel@gluster.org
>> Sent: Tuesday, March 22, 2016 1:18:18 PM
>> Subject: [Gluster-devel] How to enable DIRECT IO?
>>
>> hi,
>> In the glusterfs client, execute "mount -t glusterfs volumename -o
>> direct-io-mode=enable mountpoint".
>> Does it mean that the client reads or writes the cluster files in the way of
>> DIRECT IO? If not,how to enable DIRECT IO?
>
>This might help:
>http://www.gluster.org/pipermail/gluster-devel/2016-February/048425.html
>
>regards,
>Raghavendra
</pre></div><br><br><span title="neteasefooter"><p> </p></span>