I can only guess this to be a bug. I'd file that and probably raise it to
critical since this seems to bet key functionality broken. Have you tried this
with Cloudstack 4.2 that's nearing release, and you might be able to get it
fixed there.
Ahmad
On Jul 23, 2013, at 10:07 AM, Dinu Arateanu <dinu.arate...@gmail.com> wrote:
> Hello,
>
> I've been testing Cloudstack 4.1 with KVM/Openvswitch networking. When trying
> to add a VPC, the VR's public interface default route was not assigned
> correctly, nor the source nat. Cloudstack reports the VPC is created
> successfully, however the VR is left in an "incomplete" state. Looking
> through the agent logs I found:
>
> 2013-07-19 16:39:20,961 DEBUG [cloud.agent.Agent]
> (agentRequest-Handler-2:null) Processing command:
> com.cloud.agent.api.PlugNicCommand
> 2013-07-19 16:39:20,970 DEBUG [kvm.resource.OvsVifDriver]
> (agentRequest-Handler-2:null) plugging nic=[Nic:Public-192.168.1.68-vlan://32]
> 2013-07-19 16:39:20,970 DEBUG [kvm.resource.OvsVifDriver]
> (agentRequest-Handler-2:null) creating a vlan dev and bridge for public
> traffic per traffic label vswitch0
> 2013-07-19 16:39:21,116 DEBUG [cloud.agent.Agent]
> (agentRequest-Handler-2:null) Processing command:
> com.cloud.agent.api.routing.IpAssocVpcCommand
> 2013-07-19 16:39:21,126 DEBUG
> [resource.virtualnetwork.VirtualRoutingResource]
> (agentRequest-Handler-2:null) Executing:
> /usr/share/cloudstack-common/scripts/network/domr/router_proxy.sh
> vpc_ipassoc.sh 169.254.2.23 -A -l 192.168.1.68 -c ethnull -g 192.168.1.1 -m
> 24 -n 192.168.1.0
> 2013-07-19 16:39:29,107 DEBUG [kvm.resource.LibvirtComputingResource]
> (UgentTask-5:null) Executing:
> /usr/share/cloudstack-common/scripts/vm/network/security_group.py
> get_rule_logs_for_vms
> 2013-07-19 16:39:29,233 DEBUG [kvm.resource.LibvirtComputingResource]
> (UgentTask-5:null) Execution is successful.
> 2013-07-19 16:39:29,235 DEBUG [cloud.agent.Agent] (UgentTask-5:null) Sending
> ping: Seq 7-103: { Cmd , MgmtId: -1, via: 7, Ver: v1, Flags: 11,
> [{"PingRoutingWithNwGroupsCommand":{"newGroupStates":{},"newStates":{},"_gatewayAccessible":true,"_vnetAccessible":true,"hostType":"Routing","hostId":7,"wait":0}}]
> }
> 2013-07-19 16:39:29,243 DEBUG [cloud.agent.Agent] (Agent-Handler-5:null)
> Received response: Seq 7-103: { Ans: , MgmtId: 112938636298, via: 7, Ver:
> v1, Flags: 100010,
> [{"PingAnswer":{"_command":{"hostType":"Routing","hostId":7,"wait":0},"result":true,"wait":0}}]
> }
> 2013-07-19 16:39:38,707 DEBUG
> [resource.virtualnetwork.VirtualRoutingResource]
> (agentRequest-Handler-2:null) Execution is successful.
> 2013-07-19 16:39:38,708 DEBUG
> [resource.virtualnetwork.VirtualRoutingResource]
> (agentRequest-Handler-2:null) Device "ethnull" does not exist.
> Cannot find device "ethnull"
> Error: argument "Table_ethnull" is wrong: "table" value is invalid
>
> Error: argument "Table_ethnull" is wrong: "table" value is invalid
>
> RTNETLINK answers: No such process
>
> 2013-07-19 16:39:38,709 DEBUG [cloud.agent.Agent]
> (agentRequest-Handler-2:null) Processing command:
> com.cloud.agent.api.routing.SetSourceNatCommand
> 2013-07-19 16:39:38,719 DEBUG [kvm.resource.LibvirtComputingResource]
> (agentRequest-Handler-2:null) failing to get physical interface from
> bridgecloud0, did not find an eth*, bond*, or vlan* in
> /sys/devices/virtual/net/cloud0/brif
> 2013-07-19 16:39:38,719 DEBUG [kvm.resource.LibvirtComputingResource]
> (agentRequest-Handler-2:null) failed to get vlan id from bridge
> cloud0attached to physical interface
> 2013-07-19 16:39:38,719 DEBUG [kvm.resource.LibvirtComputingResource]
> (agentRequest-Handler-2:null) failing to get physical interface from
> bridgevswitch0, did not find an eth*, bond*, or vlan* in
> /sys/devices/virtual/net/vswitch0/brif
> 2013-07-19 16:39:38,719 DEBUG [kvm.resource.LibvirtComputingResource]
> (agentRequest-Handler-2:null) failed to get vlan id from bridge
> vswitch0attached to physical interface
> 2013-07-19 16:39:38,719 DEBUG
> [resource.virtualnetwork.VirtualRoutingResource]
> (agentRequest-Handler-2:null) Executing:
> /usr/share/cloudstack-common/scripts/network/domr/router_proxy.sh vpc_snat.sh
> 169.254.2.23 -A -l 192.168.1.68 -c eth2
> 2013-07-19 16:39:38,947 DEBUG
> [resource.virtualnetwork.VirtualRoutingResource]
> (agentRequest-Handler-2:null) Execution is successful.
> 2013-07-19 16:39:38,948 DEBUG
> [resource.virtualnetwork.VirtualRoutingResource]
> (agentRequest-Handler-2:null) iptables: Bad rule (does a matching rule exist
> in that chain?).
> iptables: No chain/target/match by that name.
>
>
> Cloudstack is configured with advanced networking. Tags for physical networks
> - vswitch0 for public & guest traffic
> - vif9 for storage traffic
> - vif8 for management traffic
>
> Openvswitch configuration:
> # ovs-vsctl show
>
> Bridge "vswitch1"
> Port "vswitch1"
> Interface "vswitch1"
> type: internal
> Port "eth1"
> Interface "eth1"
> Port "vif9"
> tag: 9
> Interface "vif9"
> type: internal
> Bridge "vswitch0"
> Port "vnet1"
> tag: 32
> Interface "vnet1"
> Port "vswitch0"
> Interface "vswitch0"
> type: internal
> Port "vif8"
> tag: 8
> Interface "vif8"
> type: internal
> Port "eth0"
> Interface "eth0"
> Bridge "cloud0"
> Port "vnet0"
> Interface "vnet0"
> Port "cloud0"
> Interface "cloud0"
> type: internal
> ovs_version: "1.9.0"
>
> Host is Ubuntu 13.04 (up-to-date) with 2 physical nics.
>
> Any idea what could be wrong?
>
> Thanks,
> --
> Dinu
>
