|
On Wed, May 14, 2003 at 03:16:02PM -0700, Dusten Splan wrote:
Hi Dusten.
> I thought that this was solved with the 2.4 kernel, because it was meant to
> be multithreaded. I'm running 2.4.20 with the newest LVS. Maybe I need to
> run 2.5 but this is a production machine. Has anyone tried LVS on 2.5?
First, as Peter said, LVS really should utilise both CPU.
As you note the 2.4 kernels are multithreaded. LVS should
take advantage of this. It definately bears further investigation.
As to the issue that Peter raised about performance. The problem
is that in multithreading the kernel a lot of spinlocks were introduced.
>From the testing that I was involved in, its seems that the overhead
in obtaining these locks is greater than the advange of having
access to a second CPU. That is in the case of using the box only
as an LVS Linux Director. If you are doing lots of other things
as well then this may not be the case.
This however, does not answer, and is not particularly relevant
to your problem. Sorry.
> I have tried the irqbalance and that doesn't seam to be helping any, Oh and
> does anyone know of a web page for that patch. I guess I just might get a
> 3GHz box and forget the hole smp thing. Here's a sample of what mpstat has
> to say about the hole thing.
I would suggest that if you are building a machine that will
act primarily as an LVS ldirectord, then a non-SMP kernel should
give you the best performance.
--
Horms
|
