[Samba] Samba + Windos XP/2k + Netbench Problem

[Dieter Stampfer]

	Herb,

fist off, thanks for your reply. I was perfectly ready to attribute the
problem to Netbench until I found that 'nt smb support = no' fixes it
for us. We've run 6 clients with 8 engines each successfully quite often
since. Clean directory structure or old files present, no more engine
failures.

Did you try with 'nt smb support = no' ? Not a single engine should die
or there's some contributing factor we're unaware of. Unless Netbench is
complete crap, I don't like the idea that more engines die with Samba
than with Win2k, even in your test.

We have these servers to test with:
Win2K       - 2P 1GHz PIII
Netware 6   - 2P 800 MHz PIII
NT4.0       - 1P 1,266 GHz PIII-S
Samba 2.2.4 - 1P 1,266 GHz PIII-S

NT4 is not an option for us but we've had similar results, i.e. all but
2 or 3 engines died. It's too early to claim that Samba beats all others
but it certainly looks a lot like it.

If you can get around to test with 'nt smb support = no' I'd be quite
interested in what you find. I guess 'nt smb support = no' is safe in
terms of keeping all other functionality the same?

We'll try again with Win2k to see if we can duplicate your results there.

	Thanks,
		Dieter

On Mon, 13 May 2002, Herb Lewis wrote:

> This may not be a samba problem per se. I just ran a test with a
> single client running 8 engines.
>
> With clean directory structure -
> Against Win2K server - 2 engines died (1 delete, 1 move)
> Against NT4.0 server - 6 engines died (3 create, 3 delete)
> Against samba 2.2.4 on IRIX - 6 engines died (1 create, 3 delete, 2 move)
>
> With old files present -
> Against a Win2K server - 2 engines died (1 create, 1 delete).
> Against NT4.0 server - 5 engines died (5 delete)
> Against samba 2.2.4 on IRIX - 6 engines died (6 delete)
>
> The servers used were as follows
> Win2K - 1P 200 MHz PentiumPro
> NT4.0 - 4P 400 MHZ Xeon
> Samba - 4P 500Mhz MIPS R14000