[Samba] File xfer speed issues with MacOS X 10.2 and 10.3

[Mike Ely]

I might - for some reason, Apple has set the default MTU to be 1500 - 
lower than other clients IIRC.  Try playing with that setting on the 
client machines (on panther go to syspref, networks, built-in ethernet, 
ethernet, and click on the "configure manually" radio button) and set 
the MTU to be a more binary-friendly value.  It might help - might not 
as well.

If anyone else has greater wisdom on this, I'm all ears.

Mike

On Nov 6, 2003, at 10:40 AM, Mark Ferlatte wrote:

> Hi all,
>
> I have a mixed environment of Windows 2000, XP, Linux, and MacOS X 
> clients
> connecting to various Samba servers.  Our network is 100Mbit switched 
> ethernet.
>
> Originally, we were using samba 2.2.8a, with the socket options of 
> TCP_NODELAY
> SO_SNDBUF=4096 SO_RCVBUF=4096.  However, the MacOS X clients (which 
> were added
> after the initial setup and testing), had horrible file xfer 
> performance;
> somewhere around 100kbit/second, while the Windows clients were much 
> faster.
>
> The first response was to upgrade to samba 3.0.0, which didn't affect 
> the
> problem one way or another.  However, setting the SNDBUF and RCVBUF to 
> 8192
> did.  With the socket buffers at 8k, all of the Macs are happy, and 
> getting
> Mbit speeds easily.
>
> Has anyone else experienced this?  Can anyone explain why server 
> socket buffers
> would have such a large effect on the client?
>
> M
> -- 
> To unsubscribe from this list go to the following URL and read the
> instructions:  http://lists.samba.org/mailman/listinfo/samba

*************Begin reminder/sig********************
In my code I always put the constant on the lhs so that the difference 
between the equality (==) and assignment (=) operator are caught by the 
compiler by accident.

if ((options == (__WCLONE|__WALL)) && (current->uid = 0))

In this case, it would make an attempted root hole more visible, as (0 
= current->uid) would not compile.
***
Ripped from an AC post at /.

---
[This E-mail scanned for viruses by Declude Virus]