Re: Network Slowdowns?

[Bill Hacker <wbh@xxxxxxxxxxxxx>]

Jamie wrote:

> Just to let anyone know..
>
> As pr. sephe's instructions I did this:
>
> line 433 in /usr/src/sys/dev/netif/xl/if_xlreg.h
>
> Changed this:
>
> #define XL_MIN_FRAMELEN 60                                                                                                   
>
> To this:
>                                                                                                                              
> #define XL_MIN_FRAMELEN 240  
>
> I don't know what that did though?
>
>
> So far, no errors but file copies are still slow. Makes me suspect the
> underrun errors weren't related?
>
> About ~ 1 mb/sec across a 100mbs ethernet connection. Correct me if I'm
> wrong, but shouldn't this be closer to 5-10 megabytes/sec, assuming 100mbs
> is "bits pr. second" and NFS had a rather large overhead? (say, 1/2 of it
> is protocol related?)

With the most-bandwidth-efficient of uncompressed protocols (never mind 'robust' 
for the moment) such as IBM bisync or RS-232 serial, you can treat a 'byte' as 
10 bits, rule of thumb.

It goes downhill from there - ethernet itself being *much* less b/w efficient 
than, for example, ARCnet/TCNS or even 100VG-Any-LAN, and TCP/IP less efficient 
on low-latency transmission paths than, for example SNA or even IPX/SPX.

File-system overhead aside, 40% overhead is probably close for most TCP/IP over 
ethernet applications, so, yes - 50% overhead should be close.

The Ethernet+TCP/IP payback, of course, is pretty decent soft-fault-tolerance 
and much easier 'universal' configuration than other optins - which is why most 
(around 100 alternatives) have faded from the scene.

IF you happen to be testing between two fs locations (NFS exported or otherwise) 
that live on the same media, you can expect to encounter very large slowdowns 
from head-positioner thrashing.

However, none of these seem to be related to the basic problem you have been 
reporting - don't be distracted by them.

The real issue still seems to be an 'intermediate level' hardware, firmware, or 
driver glitch.

> The ifconfig up/down trick doesn't seem to make a difference at this point. 
>
>
> NFS still locks up as well. (I just did a cat /dev/zero >$NFS_MOUNTED_PATH/zero) 
> and it (NFS) froze) 
>
> Wish I had used dd instead... can't seem to CTRL-C the cat process now. :-/
>
> Jamie

(one presumes our cable-plant and any routers are in good shape?)

Hmm....

What DOES happen if you aliase-up a second IP on your NIC and run 
'hardware-less' test traffic that doesn't even leave the box?

HTH,

Bill