Re: Fwd: Dom0 OOPS - causes restart

[Niraj Tolia <ntolia@gmail.com>]

On Mon, 12 Jul 2004 17:06:33 +0100, Ian Pratt <ian.pratt@cl.cam.ac.uk> wrote:
> > Sorry if this email is a duplicate but that patch seems to work. Three
> > runs of Postmark completed without problems. I will let you know in
> > case anything else turns up though.
> 
> While debugging this we spotted what looks like a network
> performance issue: We were only receiving 500Mb/s of TCP
> throughput into domain 1 from the LAN rather than the usual
> 895Mb/s (which is what you get for transmit; 128KB socket buffer
> size).
> 
> Are you seeing any odd performance numbers?
> 

Still not at the stage yet when I can start comparing. In fact, the
odds are that I might not see similar problems in my setup as my iSCSI
initiator is a hardware card (the Intel 1000 ProT) that does its own
networking outside of xen. However, I will definitely let you know if
I see something strange.

Also, if you haven't done this already and feel that it might be
useful, I can try running netperf so see what kind of bandwidth I
observe.

Niraj


> It's possible that enabling CONFIG_NETFILTER by default was a
> mistake.
> 
> It's also possible that we've introduced some other more subtle
> performance bug :-(
> 
> Ian
> 
>


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com