If I understood correctly, you are using a nested solution where the node is itself a VM.
If this is the case, you may have a security feature active on the physical host/node.
In oVirt we have the default filter which allows only a single source mac to exit the vNIC, but you disabled it.
I guess you should have something similar on the upper node.

Please note that such topics fit the user list, not devel.

Thanks,
Edy.


On Fri, Jun 30, 2017 at 1:46 PM, pengyixiang <yxpengi386@163.com> wrote:
hi, everyone

    I am a new bird in ovirt, and I transplate ovirt node to debian jessie(customed) system, it's work well now, then we create vm, use default nic1, like this:

    in vm, we ping node, it's ok;  and in node, we ping vm, it's ok too; then in node, we ping gatway or other computer in the lan, it's ok too;

but when in vm, we ping gatway or other computer in the lan, it's not ok, why is this?


>>> next is some test infomation:

1) after vm up, vnet0 is created, and use :

# brctl show ovirtmgmt

we get this:

bridge name    bridge id        STP enabled    interfaces
ovirtmgmt        8000.000c29e3037e    no        eth0
                                                                                vnet0


2) then we test network in three host in the same time

  in vm

# tcpdump -i eth0 arp

vm send request

vm not recv reply

# ping hostA


  in node(vm in)

# tcpdump -i vnet0 arp

vnet0 send request

vnet0 not recv reply


  in third host in the lan(hostA)

# tcpdump -i eth0 arp

eth0 recv request

eth0 send reply


problem like this:


why is this, do I lost some configurations ?


ovirtmgmt configuration:


vnet info:(node installed in ESXI)

   


Look forward to your reply


Have a nice day

-- 

_______________________________________________
Devel mailing list
Devel@ovirt.org
http://lists.ovirt.org/mailman/listinfo/devel