<br>Thanks all you guys for helping me. Sorry for my late reply. You gave me some very useful info. I will consider it carefully and maybe will ask your help once again :) <br><br>Many thanks <br><div class="gmail_quote">On Sat, Jan 16, 2010 at 7:02 AM, Nicolas Williams <span dir="ltr"><<a href="mailto:Nicolas.Williams@sun.com">Nicolas.Williams@sun.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div class="im">On Thu, Jan 14, 2010 at 09:49:13AM +0700, Lex wrote:<br>
</div><div class="im">> You mean lustre client? Our lustre client work as a web server for<br>
> downloading file, it's using ssd drive and the free space is quite small,<br>
> just few Gbs, but it's serving many 10GB size file quite well. If the<br>
> joining file process is on the client, how the big file can fit the poor<br>
> free space well like this ? Could you please explain it for me ?<br>
<br>
</div><div><div></div><div class="h5">First, I thought you meant by "file join" something very different than<br>
what we (Lustre team) mean by "file join". We mean this: an<br>
optimization of "cat file1 file2 ... > joined-file". I think you meant:<br>
how the illusion of a single file is created out of multiple stripes.<br>
<br>
Do correct me if my interpretation of your question was incorrect.<br>
<br>
If I understood you correctly, then there is no "physically joined"<br>
version of a file, except in the case of files with a single stripe.<br>
<br>
The illusion of "joined" stripes is presented by the client.<br>
<br>
The client need not have enough memory to store a whole file. It only<br>
needs enough for the application's needs. E.g., if you have a 1TB file<br>
and 4GB of virtual memory, then you can't read the whole file into<br>
memory, but you can read a couple of GBs at a time, do some work, then<br>
read the next couple of GBs.<br>
<br>
Nico<br>
--<br>
</div></div></blockquote></div><br>