<div dir="ltr">Thank you Chris, I'm sort of clear now.<div>In my question, stripe 0,4 means one process wants to access stripe 0 and 4 at the same time.</div><div>there is another process wants to access both stripe 0 and 2,</div>
<div>even though stripe 0, 2, 4 are in the same place (one file), <br><div class="gmail_extra">but their offsets are different, i.e., 0 and 2 are contiguous, while from 0 to 4 there is a gap.</div><div class="gmail_extra">
So my concern is, will the two processes have different I/O cost?<br>In other words, accessing 0 and 4 would take longer time than accessing 0 and 2.</div><div class="gmail_extra"><br></div><div class="gmail_extra" style>
Jaln</div><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Jun 13, 2013 at 5:23 PM, Christopher J. Morrone <span dir="ltr"><<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">In that case, it is the question part that I do not understand. :) What is "stripe 0,4", why could it be "closer" then "stripe 0,2"? In your example, 0, 2, and 4 are all in the same place.<br>
<br>
If you file is striped over 2 OSTs, then essentially what happens behind the scenes is that there are two files, one on each OST. But Lustre hides that from you, as a user. Lustre basically does modulo operations to translate a file offset from the file that it presents to the user, into which ost and offset into said ost's file to use.<br>
<br>
Does that help at all?<br>
<br>
Chris<div class="im"><br>
<br>
On 06/13/2013 02:58 PM, Jaln wrote:<br>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">
Oh, I mean there is one file, for example 6 MB, the stripe size is 1MB,<br>
and only 2 OST,<br>
then the file will be divided into 6 stripes, denoted as stripe 0,1,2,3,4,5.<br>
the distribution on the 2 OST would be stripe 0,2,4 on OST0, stripe<br>
1,3,5 on OST1.<br>
<br>
Jaln<br>
<br>
<br>
On Thu, Jun 13, 2013 at 2:54 PM, Christopher J. Morrone<br></div><div class="im">
<<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a> <mailto:<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a>>> wrote:<br>
<br>
I think you may be confused about what a stripe is in Lustre. If<br>
there are only 2 OST, then you can only stripe a file across 2.<br>
<br>
Or maybe I don't understand your terminology. I don't know what you<br>
mean by "0,4" and "0,2".<br>
<br>
<br>
On 06/13/2013 02:38 PM, Jaln wrote:<br>
<br>
if I have 6 stripes, 2 OST, using round-robin striping,<br>
stripe 0,2,4 will be on OST0,<br>
stripe 1,3,5 will be on OST1,<br>
Do you guys have any idea about what will be the difference of<br>
accessing<br>
stripe 0,4 vs stripe 0,2?<br>
stripe 0, 2 seems to be closer than 0,4, or the lustre will do<br>
some intelligent work?<br>
<br>
Jaln<br>
<br>
<br>
On Thu, Jun 13, 2013 at 10:22 AM, Christopher J. Morrone<br>
<<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a> <mailto:<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a>><br></div><div class="im">
<mailto:<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a> <mailto:<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a>>>> wrote:<br>
<br>
On 06/13/2013 05:19 AM, E.S. Rosenberg wrote:<br>
> On Thu, Jun 13, 2013 at 3:09 AM, Christopher J. Morrone<br>
> <<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a> <mailto:<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a>><br></div><div class="im">
<mailto:<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a> <mailto:<a href="mailto:morrone2@llnl.gov" target="_blank">morrone2@llnl.gov</a>>>> wrote:<br>
>> Lustre does not manage the individual disks. I sits<br>
on top of a<br>
>> filesystem, either ldiskfs(basically ext4) or zfs (as<br>
of Lustre<br>
2.4).<br>
> Is ZFS the recommended fs, or just an option?<br>
> Doesn't ZFS suffer major performance drawbacks on linux<br>
due to it<br>
> living in userspace?<br>
> Thanks,<br>
> Eli<br>
<br>
LLNL (Brian Behlendorf) ported ZFS natively to Linux. We<br>
are not using<br>
the FUSE (userspace) version. You can find it at:<br>
<br>
<a href="http://zfsonlinux.org" target="_blank">http://zfsonlinux.org</a><br>
<br>
ZFS is one of the two backend filesystem options for<br>
Lustre, as of<br>
Lustre 2.4. 2.4 is the first Lustre release that fully<br>
supports using<br>
ZFS. Here at LLNL we are using it on our newest, and<br>
largest at 55PB,<br>
filesystem.<br>
<br>
Chris<br>
<br></div>
______________________________<u></u>___________________<br>
Lustre-discuss mailing list<br>
Lustre-discuss@lists.lustre.__<u></u>org<br>
<mailto:<a href="mailto:Lustre-discuss@lists.lustre.org" target="_blank">Lustre-discuss@lists.<u></u>lustre.org</a>><br>
<mailto:<a href="mailto:Lustre-discuss@lists." target="_blank">Lustre-discuss@lists.</a>_<u></u>_<a href="http://lustre.org" target="_blank">lustre.org</a><br>
<mailto:<a href="mailto:Lustre-discuss@lists.lustre.org" target="_blank">Lustre-discuss@lists.<u></u>lustre.org</a>>><br>
<br>
<a href="http://lists.lustre.org/__mailman/listinfo/lustre-__discuss" target="_blank">http://lists.lustre.org/__<u></u>mailman/listinfo/lustre-__<u></u>discuss</a><div class="im"><br>
<<a href="http://lists.lustre.org/mailman/listinfo/lustre-discuss" target="_blank">http://lists.lustre.org/<u></u>mailman/listinfo/lustre-<u></u>discuss</a>><br>
<br>
<br>
<br>
<br>
--<br>
<br>
Genius only means hard-working all one's life<br>
<br>
<br>
<br>
<br>
<br>
--<br>
<br>
Genius only means hard-working all one's life<br>
<br>
</div></blockquote>
<br>
</blockquote></div><br><br clear="all"><div><br></div>-- <br><pre style="margin:0in;font-family:Arial,Helvetica,sans-serif;font-size:large" lang="en-US" align="left"><font><font face="comic sans ms,sans-serif"><font color="#666666"><span>Genius only means hard-working all one's life</span></font></font></font></pre>
</div></div></div>