>
> I recently upgraded one of our servers from a Pentium 120 to an AMD K6-2
> 400, keeping the network cards (3c59x and Tlan) and most of the other
> hardware, and leaving the kernel (2.0.35), etc., intact. We're now
> experiencing a strange problem with ftp service - most bulk transfers seem
> to stall after a few K and stay that way. The tcpdumps I've done remotely
> show the remote end not receiving segments and repeatedly acking an old
> segment. Users report similar problems with bulk web transfers. Upgrading
> to 2.0.38 didn't help. Swapping network interfaces didn't either.
I have similar rpoblems, we use Cisco catalysts, I am suspecting a
combination of factors here.
I seem to get better transferrates when I switch to 10Mbit half duplex or
use a 3com stackbuilder as a switch.
I haven't isolated the problem yet, but I just put in an RFH for testing
equipment.
(Request For Hardware)
I'll pick that up today I hope and see if I can find time and clues about
this.
> Everything else appears to be normal, including interactive network
> sessions and transfers to the system.
Did you try FTP, NFS, CPIO over rsh, HTTP ?
I intend to do both UDP and TCP transfers and give it a good beating to
see where bandwidth is being lost.
> Any suggestions short of moving the system to 2.2? (Which I hope to do
> before the end of the month, but may not have time for.)
If it's the same problem I am having, 2.2 will not help.
I intend to try some other OSes as well, to see if it is the linux driver
in combination with other things, or a compatibility issue with our
switches or what.
---------------------------draw-conclusion-here------------------------
Homme R. Bitter *NIX admin, BOFH, MCSE, parttime divine entity.
(S) This signature is swingware, I let you use mine if you let me use
yours.....
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/