[one-users] vnet interfaces using open vswitch

[Carlos Jiménez]

Hi all,

We're running OpenNebula 3.8.3 with Open vSwitch and we've found out an 
issue. Once the frontend and the host are started, the VMs appear in 
Pending state and move to Prolog and again back to Pending states.

This is the output of oned.log:

/Tue Feb  5 11:30:04 2013 [DiM][D]: Deploying VM 0
Tue Feb  5 11:30:04 2013 [ReM][D]: Req:5360 UID:0 VirtualMachineDeploy 
result SUCCESS, 0
Tue Feb  5 11:30:07 2013 [TM][D]: Message received: LOG I 0 clone: 
Cloning /var/lib/one/datastores/1/d76d1fd89f175e1027f8506978165c03 in 
host1:/var/lib/one//datastores/0/0/disk.0
Tue Feb  5 11:30:07 2013 [TM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:30:07 2013 [TM][D]: Message received: LOG I 0 ln: Linking 
/var/lib/one/datastores/1/923331c1aeb5a587dd428d0b8607ff29 in 
host1:/var/lib/one//datastores/0/0/disk.1
Tue Feb  5 11:30:07 2013 [TM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:30:07 2013 [TM][D]: Message received: TRANSFER SUCCESS 0 -
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG I 0 
Successfully execute network driver operation: pre.
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG I 0 Command 
execution fail: cat << EOT | /var/tmp/one/vmm/kvm/deploy 
/var/lib/one//datastores/0/0/deployment.24 host1 0 host1
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG I 0 error: 
Failed to create domain from /var/lib/one//datastores/0/0/deployment.24
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG I 0 error: 
Unable to add bridge vbr1 port vnet0: Invalid argument
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG E 0 Could not 
create domain from /var/lib/one//datastores/0/0/deployment.24
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG I 0 ExitCode: 255
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: LOG I 0 Failed to 
execute virtualization driver operation: deploy.
Tue Feb  5 11:30:08 2013 [VMM][D]: Message received: DEPLOY FAILURE 0 
Could not create domain from /var/lib/one//datastores/0/0/deployment.24/

We've realised that one tries to create a vnetx but that vnet interface 
is already into the Open vSwitch database, so it is unable to introduce 
that interface and therefore to create the VM. This is the output of the 
openvswitch:
/#ovs-vsctl show
6725e67a-3af1-4fdf-9dfe-f606d09918a8
     Bridge "vbr1"
         Port "bond0"
             Interface "bond0"
         Port "vbr1"
             Interface "vbr1"
                 type: internal
     ovs_version: "1.4.3"/

We've managed to solve it manually deleting those interfaces into the 
open vswitch database and immeditely one has been able to create the VMs.
This is the output:

/Tue Feb  5 11:31:37 2013 [TM][D]: Message received: LOG I 0 clone: 
Cloning /var/lib/one/datastores/1/d76d1fd89f175e1027f8506978165c03 in 
host1:/var/lib/one//datastores/0/0/disk.0
Tue Feb  5 11:31:37 2013 [TM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:31:37 2013 [TM][D]: Message received: LOG I 0 ln: Linking 
/var/lib/one/datastores/1/923331c1aeb5a587dd428d0b8607ff29 in 
host1:/var/lib/one//datastores/0/0/disk.1
Tue Feb  5 11:31:37 2013 [TM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:31:37 2013 [TM][D]: Message received: TRANSFER SUCCESS 0 -
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 
Successfully execute network driver operation: pre.
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 
Successfully execute virtualization driver operation: deploy.
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 post: 
Executed "sudo /usr/bin/ovs-ofctl add-flow vbr1 
in_port=2,dl_src=02:00:c0:a8:0f:64,priority=40000,actions=normal".
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 post: 
Executed "sudo /usr/bin/ovs-ofctl add-flow vbr1 
in_port=2,priority=39000,actions=drop".
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 ExitCode: 0
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: LOG I 0 
Successfully execute network driver operation: post.
Tue Feb  5 11:31:38 2013 [VMM][D]: Message received: DEPLOY SUCCESS 0 one-0/

Is there any way to manage it? We've thought on an script to 
automatically check it everytime we restart the servers, but perhaps 
there is already a better way we unknow.


Thanks in advance,

Carlos.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20130205/808e00d5/attachment-0001.htm>