Re: 2.5.60 NFS FSX

Dave Jones (davej@codemonkey.org.uk)
Thu, 13 Feb 2003 21:30:48 +0000


On Thu, Feb 13, 2003 at 10:12:58PM +0100, Trond Myklebust wrote:

> > There seems to be some garbage at the end of the capture. I'm
> > not sure why, but it seems tcpdump does that sometimes.
> Does it do that on 2.4.x? I've certainly never seen that happen on a
> stable kernel.

I'll try in a few minutes. It doesn't always do it in 2.5, just like
once every dozen or so captures. NIC is a 3com 3c905C-TX/TX-M [Tornado]
in the client, and an Intel 82801BA/BAM/CA/CAM in the server, both
boxes connected through a cheapo 100MB switch. Nothing special.

> In fact ethereal reports
>
> Message: pcap: File has 1045168887-byte packet, bigger than maximum of 65535
> No wonder we see bizarre crashes...

That did seem odd yes. I put it down to a tcpdump bug.
But if you think thats whats triggering it...

> There are several other odd features in this tcpdump. Random UDP
> packets from the server to the client with a junk payload (usually
> consisting of a load of zeros) appear to be pretty frequent.

Wacky. Interested netdev parties can find the ~2MB tcpdump output at
http://www.codemonkey.org.uk/cruft/tcp-trond.bz2
(please don't hammer that server with requests for this unless you
really do want/need it - the mind boggles at how many hits the nfs
fsx dumps got in the past -- I'm sure we don't have *that* many
interested NFS developers 8-).

> Is this against a 2.5.x server? If so, could you try against a 2.4.x
> or a non-linux server?

This is a 2.4.21pre3 server. No non-linux servers to try against
right now..

Dave

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/