Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 6 Jul 2013 12:34:41 -0700
From:      Martin Alejandro Paredes Sanchez <mapsware@prodigy.net.mx>
To:        freebsd-questions@freebsd.org
Subject:   Re: Possibly OT: NFS vs SMB performance
Message-ID:  <201307061234.41962.mapsware@prodigy.net.mx>
In-Reply-To: <51D7DB83.4060809@netfence.it>
References:  <51D6F1E4.4090001@netfence.it> <669058E9-E663-424E-94A6-29D81757C580@elde.net> <51D7DB83.4060809@netfence.it>

next in thread | previous in thread | raw e-mail | index | archive | help
On Saturday 06 July 2013 01:55:31 Andrea Venturoli wrote:
> On 07/05/13 20:42, Terje Elde wrote:
> > On 5. juli 2013, at 18:18, Andrea Venturoli <ml@netfence.it> wrote:
> >> Is this normal in your experience?
> >
> > Did you do them in that order, or did you do the smb (slow) one first?
> >
> > If the slow was first, I'm thinking caching on the server could be a
> > major factor.
>
> Yesterday I did four test:
> _ SMB find resulting in over 10 minutes first time;
> _ SMB find resulting in nearly 10 minutes second time;
> _ NFS find resulting in a little over 1 minute first time;
> _ NFS find resulting in a little less than 1 minute second time.
>
>
> Today I tried again in reverse order:
> _ NFS find took 3 minutes;
> _ NFS find again took 21 seconds;
> _ SMB find took over 9 minutes;
> _ SMB find again took again over 9 minutes.
>
> So, while caching plays a role, it just isn't it.
> The server was possibly doing other things, so the above figures might
> not be that correct; however a difference in the magnitude order is just
> too big (and deterministic) to be considered random noise.
>

the problem may be high log level for Samba

You should read this

http://www.hob-techtalk.com/2009/03/09/nfs-vs-cifs-aka-smb



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201307061234.41962.mapsware>