<div dir="ltr"><div>hola jeff: im not sure wether your volfile command is complimentary, or alternative , to my simple and easy "mount with entry-timeout=0" option. <br><br></div>tom : Im not sure , lets wait for jeff, he's the hardcore gluster consistency expert. <br>
<br>im just a user :)<br><div><br><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Wed, Feb 12, 2014 at 5:23 PM, Tom Munro Glass <span dir="ltr"><<a href="mailto:tom@tmgcon.com" target="_blank">tom@tmgcon.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On 02/13/2014 10:24 AM, Jay Vyas wrote:<br>
> For vanilla apps that are doing stuff in gluster, you normally do it<br>
> through a fuse mount.<br>
><br>
> mount -t glusterfs localhost:HadoopVol /mnt/glusterfs<br>
><br>
> But in your case, you might want to do some strict consistency settings to<br>
> make it atomic:<br>
><br>
> mount -t glusterfs localhost:HadoopVol -o<br>
> entry-timeout=0,attribute-timeout=0/mnt/glusterfs<br>
><br>
> This will make sure that everything is refreshed when you look up files.<br>
> This strategy has solved our eventual consistency requirements for the<br>
> hadoop plugin.<br>
><br>
</div></div>Are you saying that with these mount options I can just write files<br>
directly without using flock or renaming a temporary file, and that<br>
other processes trying to read the file will always see a complete and<br>
consistent view of the file?<br>
<span class="HOEnZb"><font color="#888888"><br>
Tom<br>
<br>
</font></span></blockquote></div><br><br clear="all"><br>-- <br>Jay Vyas<br><a href="http://jayunit100.blogspot.com" target="_blank">http://jayunit100.blogspot.com</a>
</div>