Can you tell me what you mean by a vnet number? I do not know that term. On Wed, Sep 19, 2012 at 09:40:41AM +0200, r po wrote: > > Hi, > > Thanks Ben for you answers on the bonding. It's really helpfull. > > I have another question but not on the same part. I'd like to supervise my > hypervisor (which hold my VM) > with a cacti-like. It's works fine and i see all vnet, bridge and fake-bridge > associate on the hypervisor. > > But there's the problems : is it possible to set a vnet number to a VM or > associate an alias ? > i explain, when i stop many VM and reboot it but not in the same order, the > vnet number will change (for each VM) > so i got to reconfigure correctly my cacti-like for specify which vnet is for > which VM and so which graph is for which VM. (or i must connect to hypervisor > see which vnet is for which VM, it's not pratical .. ) > > it's possible to do this ? or implement ? > > Thank's > > Aeron > > > Date: Tue, 18 Sep 2012 10:18:30 -0700 > > From: [email protected] > > To: [email protected] > > CC: [email protected] > > Subject: Re: [ovs-discuss] mode_bond balance-tcp and question > > > > There is definitely a difference. If you use balance-slb, then > > streams from a particular host (that is, all streams, if you are not > > using VMs) will use the same interface. If you use balance-tcp, then > > streams from a particular host will be spread (statistically) across > > the available interfaces. > > > > On Tue, Sep 18, 2012 at 07:20:08PM +0200, r po wrote: > > > Ok ! Tanks a lot for this quickly answer that conforte me in my config. > > > Can you explain me if there are à real difference on balance TCP and > > > balance slb > > > If the choice is mostly random ? > > > > > > Thanks again ! > > > > > > > Date: Tue, 18 Sep 2012 09:56:32 -0700 > > > > From: [email protected] > > > > To: [email protected] > > > > CC: [email protected] > > > > Subject: Re: [ovs-discuss] mode_bond balance-tcp and question > > > > > > > > On Tue, Sep 18, 2012 at 04:35:08PM +0200, r po wrote: > > > > > In the documentation i read, balance-tcp : balances among interface > > > > > based on L2, L3 and L4 and MAC and IP. > > > > > > > > > > But when i'm try to test the troughput (i used iperf) i don't up to > > > > > 2Go on communication between two same IP witch two différent TCP port > > > > > : > > > > > > > > > > here is my test : > > > > > > > > > > on ovs 1 : (i configured a fake bridge with IP, work well) > > > > > > > > > > iperf -c 10.10.10.10 -p 1000 > > > > > > > > > > (on other console but always in ovs1 ) > > > > > > > > > > iperf -c 10.10.10.10 -p 7000 > > > > > > > > > > on ovs2 (in two different console) > > > > > > > > > > iperf -s -p 1000 > > > > > iperf -s -p 7000 > > > > > > > > > > i got about 500Mbps on each transfert. > > > > > > > > The initial choice of interface for a given stream is essentially > > > > random. Every 10 seconds, streams get rebalanced. If you have > > > > exactly 2 streams, and each one lasts only a few seconds, whether you > > > > see a benefit is a matter of luck. > > > >
> _______________________________________________ > discuss mailing list > [email protected] > http://openvswitch.org/mailman/listinfo/discuss _______________________________________________ discuss mailing list [email protected] http://openvswitch.org/mailman/listinfo/discuss
