[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[no subject]



Mike

-----Original Message-----
From: Christopher Fowler <cfowler at outpostsentinel.com>
Sent: Sep 12, 2004 9:49 AM
To: Stephan Uphoff <ups at tree.com>
Cc: Atlanta Linux Enthusiasts <ale at ale.org>
Subject: Re: [ale] Fast NFS

I think it may be related to the ARM board.  It has only 32mb of mem and
that is not much when compiling glibc.

Here is kernel log:

__alloc_pages: 0-order allocation failed (gfp=0x92/0)
__alloc_pages: 0-order allocation failed (gfp=0x92/0)
__alloc_pages: 0-order allocation failed (gfp=0x92/0)
__alloc_pages: 0-order allocation failed (gfp=0x92/0)
__alloc_pages: 0-order allocation failed (gfp=0x92/0)

Which tells me we are running out of memeory.

On Sat, 2004-09-11 at 23:54, Stephan Uphoff wrote:
> On Sat, 2004-09-11 at 23:12, Christopher Fowler wrote:
> > Funmy.
> > 
> > I can't imagine a busty load.  I'm on a 100mbs switch and all I'm doing
> > on the ARM is compiling GCC.  What else could be going on that the NFS
> > server does not respond?  It is only serving up one share to one
> > machine.  It is doing nothing else.
> 
> Are you using NFS2 or NFS3 ?
> 
> Have you tried to reduce the write/read size all the way down to 1024 ?
> 
> Some network cards are just overwhelmed when a packet train for
> a big NFS request/reply is coming in.
> 
> What NIC are you using on the server?
> 
> How big is the receive queue on your arm board? 
> Can the board get enough memory to replace the buffers used
> by the received frames?
> 
> As always a network sniffer is your best friend.
> 
> 
> 	Stephan
> 
> 
> > 
> > 
> > On Sat, 2004-09-11 at 21:29, Pete Hardie wrote:
> > > On Sat, 11 Sep 2004 17:32:58 -0400, Stephan Uphoff <ups at tree.com> wrote:
> > > > My guess is that IP fragmentation issues (packet loss under load, busty
> > > > load .. ) are killing you. (You are using UDP right ?)
> > >
> > > "Busty load"???  When did they extend NFS to recognize porn? :->
> 
> NFS 6 ?
> 
> Not that I know of - but NFS 4 gained non persistent file handles for
> the MS crowd :-( .... so they are definitely trying for mass appeal :-)
> 
> 
> > > _______________________________________________
> > > Ale mailing list
> > > Ale at ale.org
&gt; &gt; &gt; <a  rel="nofollow" href="http://www.ale.org/mailman/listinfo/ale";>http://www.ale.org/mailman/listinfo/ale</a>
&gt; &gt; 
&gt; &gt; _______________________________________________
&gt; &gt; Ale mailing list
&gt; &gt; Ale at ale.org
&gt; &gt; <a  rel="nofollow" href="http://www.ale.org/mailman/listinfo/ale";>http://www.ale.org/mailman/listinfo/ale</a>
&gt; &gt; 
&gt; &gt; 

_______________________________________________
Ale mailing list
Ale at ale.org
<a  rel="nofollow" href="http://www.ale.org/mailman/listinfo/ale";>http://www.ale.org/mailman/listinfo/ale</a>




</pre>
<!--X-Body-of-Message-End-->
<!--X-MsgBody-End-->
<!--X-Follow-Ups-->
<hr>
<ul><li><strong>Follow-Ups</strong>:
<ul>
<li><strong><a name="00386" href="msg00386.html">[ale] Fast NFS</a></strong>
<ul><li><em>From:</em> cfowler at outpostsentinel.com (Christopher Fowler)</li></ul></li>
</ul></li></ul>
<!--X-Follow-Ups-End-->
<!--X-References-->
<!--X-References-End-->
<!--X-BotPNI-->
<ul>
<li>Prev by Date:
<strong><a href="msg00380.html">[ale] Help! moving a Red Hat 8 drive to a pre-existing XP system</a></strong>
</li>
<li>Next by Date:
<strong><a href="msg00382.html">[ale] Fast NFS</a></strong>
</li>
<li>Previous by thread:
<strong><a href="msg00377.html">[ale] Fast NFS</a></strong>
</li>
<li>Next by thread:
<strong><a href="msg00386.html">[ale] Fast NFS</a></strong>
</li>
<li>Index(es):
<ul>
<li><a href="maillist.html#00381"><strong>Date</strong></a></li>
<li><a href="threads.html#00381"><strong>Thread</strong></a></li>
</ul>
</li>
</ul>

<!--X-BotPNI-End-->
<!--X-User-Footer-->
<!--X-User-Footer-End-->
</body>
</html>