Use and development of the Network File System (NFS) components of Linux ()

Edward Roper | 1 Jul 2003 01:18

RE: close() / nfs commit performance

Fair enough, I've read the faq and understand your argument. However I
would really be curious as to why local write benchmarks (on the server
itself) with calling close() and using a test-file of 2GB (twice my ram)
and a record size of 8192 I obtain a write throughput of 24,739KB/Second
whereas over NFS I obtain only 4,924KB/Second on a 500mb file with 8192
record size?

I realize that plopping a pci-nvramish device (http://www.umem.com/) and
setting data=journal with my ext3 journal residing on the nvram would
indeed offer a performance increase. I don't see where it's going to be
a 5x increase however when I don't seem to be bottle necking at
direct-attach-disk-i/o.

Perhaps local close() calls don't actually commit to disk?

Watching the machines leds (albeit a very primitive method) doesn't show
constant harddisk action while I'm supposedly waiting on a physical disk
commit. I get a brief flash of activity once every 30 seconds or so
instead across the array.

I plan on purchasing the nvram but I'm just making sure it's really
going to make the difference that I'm looking for. 

On Mon, 2003-06-30 at 15:39, Lever, Charles wrote:
> ed-
> 
> read the FAQ:  http://nfs.sourceforge.net/
> 
> close(2) on an NFS file means flush all outstanding
> writes.  benchmarking without the close() means you're

(Continue reading)

Trond Myklebust | 1 Jul 2003 02:34

Picon

Picon

Re: close() / nfs commit performance

>>>>> " " == Edward Roper <eroper <at> wanfear.com> writes:

     > Perhaps local close() calls don't actually commit to disk?

man 2 close
   (and read the NOTES)

and then

man 2 fsync

Cheers,
  Trond

-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
NFS maillist  -  NFS <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

raysrv05 | 1 Jul 2003 07:51

Picon

Report to Recipient(s)

Incident Information:-

Originator: nfs-admin <at> lists.sourceforge.net
Recipients: <nfs <at> lists.sourceforge.net>, jd <at> ts.ray.fi
Subject:    [NFS] Re: Movie

The file your_details.zip (details.pif) you received was infected with the
W32/Sobig.e <at> MM virus and was deleted.

-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
NFS maillist  -  NFS <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Trond Myklebust | 1 Jul 2003 11:39

Picon

Picon

Re: [PATCH 1/4] Optimize NFS open() calls by means of 'intents'...


     > Can we call the union something descriptive (eg "intent")
     > rather than "u"?

Sure. I'm fine with substutiting 'intent' for 'u' in the final version.

Any other comments/objections/... before I push this on to Linus?

Cheers,
  Trond
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo <at> vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

justin case | 1 Jul 2003 13:42

Picon

server locks up totally

hi,

i actually sent this message yesterday already, but it seems, that it didn't
get through to the list due to my account not being activated already. here
it is again:

i build up a diskless-system  using etherboot and root on an nfs-share. both
systems are redhat9 with 2.4.20 kernels.

so far everything works well now but strangely the nfs-server tends to lock
up out of the blue. this is not reconstructable and just happens sooner or
later.

since it always locks up the entire machine(!) the nfs server runs on i
can't get any error-message and slowly i am a little desperate here..

i don't use any automouter and my /etc/exports file only contains:

tftpboot/gate01 192.168.1.20(rw,no_root_squash)

i can reboot the server and the client starts working again the second the
nfs-share is available again, so the client-side seems to be ok.

sorry for coming up with such little info  :(
hopefully anybody has an idea..

my regards
/jc

-------------------------------------------------------

(Continue reading)

Edward Roper | 1 Jul 2003 21:40

Re: close() / nfs commit performance

So in theory running nfs async is no more dangerous than using a local
app that calls close() without fsync()?

On Mon, 2003-06-30 at 17:34, Trond Myklebust wrote:
> >>>>> " " == Edward Roper <eroper <at> wanfear.com> writes:
> 
>      > Perhaps local close() calls don't actually commit to disk?
> 
> man 2 close
>    (and read the NOTES)
> 
> and then
> 
> man 2 fsync
> 
> Cheers,
>   Trond

-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
NFS maillist  -  NFS <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Edward Roper | 1 Jul 2003 21:55

Re: close() / nfs commit performance

Alright so rerunning iozone with the flush option which calls fsync() my
new number on the 2GB file is: 16,797kB/second vs. the 4,924kB/second on
the Linux server and the 31,757kB/second on the netapp. I'm still not
understanding where the flush performance hit to the tune of
~12,000kB/second is coming from? I would assume that with nvram I would
have a greater fsync() local performance than the 16,797kB/second?

On Mon, 2003-06-30 at 17:34, Trond Myklebust wrote:

> man 2 fsync
> 
> Cheers,
>   Trond

-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
NFS maillist  -  NFS <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Edward Roper | 1 Jul 2003 23:36

Re: close() / nfs commit performance

And one last number: data=journal with a ramdisk (trd) journal device
leads to only 8,782KB/Second. (via tcp nfs 500mb filesize 8192 record
size) vs. 4,924kB/Second. Quite an improvement, but still not
16,797kB/Second.

On Tue, 2003-07-01 at 12:55, Edward Roper wrote:
> Alright so rerunning iozone with the flush option which calls fsync() my
> new number on the 2GB file is: 16,797kB/second vs. the 4,924kB/second on
> the Linux server and the 31,757kB/second on the netapp. I'm still not
> understanding where the flush performance hit to the tune of
> ~12,000kB/second is coming from? I would assume that with nvram I would
> have a greater fsync() local performance than the 16,797kB/second?
> 
> On Mon, 2003-06-30 at 17:34, Trond Myklebust wrote:
> 
> > man 2 fsync
> > 
> > Cheers,
> >   Trond

-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
NFS maillist  -  NFS <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

(Continue reading)

raysrv05 | 2 Jul 2003 04:05

Picon

Report to Recipient(s)

Incident Information:-

Originator: nfs-admin <at> lists.sourceforge.net
Recipients: <nfs <at> lists.sourceforge.net>, jd <at> ts.ray.fi
Subject:    [NFS] Re: Movie

The file your_details.zip (details.pif) you received was infected with the
W32/Sobig.e <at> MM virus and was deleted.

-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
NFS maillist  -  NFS <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

raysrv05 | 2 Jul 2003 08:39

Picon

Report to Recipient(s)

Incident Information:-

Originator: nfs-admin <at> lists.sourceforge.net
Recipients: <nfs <at> lists.sourceforge.net>, jd <at> ts.ray.fi
Subject:    [NFS] Re: Application

The file your_details.zip (details.pif) you received was infected with the
W32/Sobig.e <at> MM virus and was deleted.

-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
NFS maillist  -  NFS <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Group

gmane.linux.nfs.

Project Web Page

Use and development of the Network File System (NFS) components of Linux ()

Search Archive

Page

1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 18

Articles in period: 177

July 2003

Language

Change language

Options

Current view: Threads only / Showing only 20 lines / Not hiding cited text.
Change to All messages, whole messages, or hide cited text.

Post a message
NNTP Newsgroup
Classic Gmane web interface
XML

XML

RSS Feed
List Information

About Gmane

Recent entries

[PATCH v2] nfsd: avoid undefined signed overflow (17 May 2013)
[PATCH] nfsd: avoid undefined signed overflow (17 May 2013)
Is the code stateid_generation_after()in legal C? (17 May 2013)
NFS loop on 3.4.39 (17 May 2013)
[PATCH] clnt_vc_control: fix VERS/PROG size (16 May 2013)
Froe e71bf1d708e1294b3bae64d04f03228b3625f2a3 Mon Sep 17 00:00:00 2001 (16 May 2013)
SUNRPC: Convert auth_gss pipe detection to work in namespaces (16 May 2013)
Linux NFS town hall meeting 05/16/13 (Cancelled) (16 May 2013)
[PATCH] NFS: Fix SETCLIENTID fallback if GSS is not available (16 May 2013)
see comments nfs_fs.h:seqid_mutating_error() (15 May 2013)
[PATCH 0/3] Speed up detection of whether or not rpc.gssd is running (15 May 2013)
[PATCH] NFSD: Don't give out read delegations on exclusive creates (15 May 2013)
Some per-machine NFS export options are broken by 11ba3b1e01b67b7d19f2 (15 May 2013)
(unknown) (14 May 2013)
[PATCH] sunrpc: server back channel needs no rpcbind method (14 May 2013)
[PATCH] NFS: Fix security flavor negotiation with legacy binary mounts (14 May 2013)
Where in the server code is fsinfo rtpref calculated? (14 May 2013)
Hangs when do ls on mount directory (14 May 2013)
Possible bug introduced by: NFS: Use server-recommended security flavo (13 May 2013)
Looping NFSv4 traffic (13 May 2013)
[PATCH 00/13] lnfs: linux-3.10-rc1 release (13 May 2013)

Archives

301	May 2013
446	April 2013
426	March 2013
427	February 2013
443	January 2013
263	December 2012
468	November 2012
578	October 2012
350	September 2012
536	August 2012
402	July 2012
363	June 2012
819	May 2012
729	April 2012
871	March 2012
657	February 2012
543	January 2012
518	December 2011
625	November 2011
408	October 2011
596	September 2011
537	August 2011
659	July 2011
828	June 2011
757	May 2011
410	April 2011
631	March 2011
529	February 2011
557	January 2011
703	December 2010
526	November 2010
704	October 2010
894	September 2010
691	August 2010
526	July 2010
680	June 2010
623	May 2010
508	April 2010
392	March 2010
351	February 2010
493	January 2010
370	December 2009
328	November 2009
274	October 2009
439	September 2009
655	August 2009
242	July 2009
716	June 2009
484	May 2009
866	April 2009
868	March 2009
262	February 2009
497	January 2009
494	December 2008
542	November 2008
470	October 2008
448	September 2008
445	August 2008
345	July 2008
650	June 2008
470	May 2008
541	April 2008
356	March 2008
483	February 2008
781	January 2008
382	December 2007
571	November 2007
551	October 2007
594	September 2007
610	August 2007
454	July 2007
234	June 2007
442	May 2007
395	April 2007
429	March 2007
246	February 2007
223	January 2007
251	December 2006
319	November 2006
274	October 2006
180	September 2006
292	August 2006
350	July 2006
227	June 2006
154	May 2006
173	April 2006
175	March 2006
149	February 2006
151	January 2006
175	December 2005
146	November 2005
176	October 2005
108	September 2005
77	August 2005
202	July 2005
179	June 2005
127	May 2005
151	April 2005
392	March 2005
213	February 2005
194	January 2005
194	December 2004
166	November 2004
176	October 2004
237	September 2004
244	August 2004
105	July 2004
136	June 2004
215	May 2004
180	April 2004
191	March 2004
231	February 2004
118	January 2004
109	December 2003
124	November 2003
133	October 2003
229	September 2003
175	August 2003
177	July 2003
214	June 2003
228	May 2003
150	April 2003
128	March 2003
146	February 2003
207	January 2003
108	December 2002
198	November 2002
367	October 2002
255	September 2002
303	August 2002
150	July 2002
93	June 2002
163	May 2002
192	April 2002
96	March 2002
2	December 1999

Design

Zawodny | Right Menu | Left Menu

Your Own Design

Paste the URL of your CSS below. Download CSS template