Hi all, I have problem reaching 1Gbps network speed with dom0 / domU (paravirtualized) on my centos5 installations. The fastest speed I measured in domU was about 800/600 receive/transmit Mbps, dom0 being a bit slower. I'm using bridging (eeither with default xen/centos network-bridge script or with my own scripts assigning vlans to bridges). Is anyone experiencing same network slowdown? Or could it be just some configuration issue on my side? Thanks, david.kostal at gmail.com ----+
I have not had the opportunity to test, but it sounds like expected behavior due how I/O is handled within Xen. It's a know fact that disc I/O is slow so I would expect the same for network I/O. - Nicolas On Nov 23, 2007 2:02 PM, David Kostal <david.kostal at gmail.com> wrote:> Hi all, > I have problem reaching 1Gbps network speed with dom0 / domU > (paravirtualized) on my centos5 installations. > > The fastest speed I measured in domU was about 800/600 > receive/transmit Mbps, dom0 being a bit slower. I'm using bridging > (eeither with default xen/centos network-bridge script or with my own > scripts assigning vlans to bridges). > > Is anyone experiencing same network slowdown? > Or could it be just some configuration issue on my side? > > Thanks, > > david.kostal at gmail.com > ----+ > _______________________________________________ > CentOS-virt mailing list > CentOS-virt at centos.org > http://lists.centos.org/mailman/listinfo/centos-virt >
David Kostal wrote:> Hi all, > I have problem reaching 1Gbps network speed with dom0 / domU > (paravirtualized) on my centos5 installations. > > The fastest speed I measured in domU was about 800/600 > receive/transmit Mbps, dom0 being a bit slower. I'm using bridging > (eeither with default xen/centos network-bridge script or with my own > scripts assigning vlans to bridges). > > Is anyone experiencing same network slowdown? > Or could it be just some configuration issue on my side?What do you get on the same hardware running a non-Xen kernel? What does the test look like? You also have to consider that Ethernet is not 100% efficient. -- Christopher G. Stach II