Hi Jack,
On 01/13/12 13:00, Jack Vogel wrote:
Opps, i meant kern.ipc.nmb* of course...
This was precisely the right place to tweak, thanks!
Iordan
___
freebsd-bugs@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-bugs
To un
Its a system resource thing, when you increase to a 9K mtu the driver will
use 9K
mbuf clusters, and with the queues and rings you have its unable to get
enough.
Very important to look at kern.ipc.nbm*
In your case nmbjumbo9 is the pool, if you had 4K it would those clusters.
Change the defaults
Opps, i meant kern.ipc.nmb* of course...
Jack
On Fri, Jan 13, 2012 at 9:58 AM, Jack Vogel wrote:
> Its a system resource thing, when you increase to a 9K mtu the driver will
> use 9K
> mbuf clusters, and with the queues and rings you have its unable to get
> enough.
>
> Very important to look
Hello,
We are testing FreeBSD 9.0-RELEASE on a pair of machines with 10GBASE-T
Intel x520-t2 adapters interconnected with a straight-through cat6 rj45
cable. With mtu 1500, we are achieving aproximately 4 gigabits/s both
directions as demonstrated with iperf. We wanted to test with mtu 9000
t
