[one-users] Using RAW attribute section for Xen

[Ruben S. Montero]

Hi

As shown in your email, Xen uses a native driver (not thourgh
libvirt). You need to add those statements that would be added to a
Xen domain config file.

You can refer to Xen documentation to check if that particular feature
is supported, and how to express it in a Xen file. Then simply add it
in a RAW attribute as in your example.

Cheers

Ruben

On Wed, Sep 5, 2012 at 8:56 AM, Yasz Yassin <yaszrina at gmail.com> wrote:
> Hi All,
>
> I want to attach the GPU during VM deployment using Xen hypervisor. I did a
> dump XML on the working VM with GPU attached and put the attributes in the
> RAW section of my VM template as below: (i am using opennebula 3.6)
> ------------------------------------
> RAW=[
>    TYPE=xen,
>    data = "<devices><hostdev
> mode=\"subsystem\",type=\"pci\",managed=\"no\"><source><address
> domain=\"0x0000\",bus=\"0x03\",slot=\"0x00\",function=\"0x0\"/></source></hostdev></devices>"
> ]
> ----------------------------------------
>
> However, this is not successful and the error log is:
> --------------------------------------------
> Wed Sep  5 14:42:30 2012 [DiM][I]: New VM state is ACTIVE.
> Wed Sep  5 14:42:30 2012 [LCM][I]: New VM state is PROLOG.
> Wed Sep  5 14:42:31 2012 [TM][I]: clone: Cloning
> ../../1/a8caf58f4964046e788b94f35c6331ec in <host
> ip>:/srv/cloud/one/var//datastores/0/80/disk.0
> Wed Sep  5 14:42:31 2012 [TM][I]: ExitCode: 0
> Wed Sep  5 14:42:32 2012 [TM][I]: context: Generating context block device
> at 10.1.32.201:/srv/cloud/one/var//datastores/0/80/disk.1
> Wed Sep  5 14:42:32 2012 [TM][I]: ExitCode: 0
> Wed Sep  5 14:42:32 2012 [LCM][I]: New VM state is BOOT
> Wed Sep  5 14:42:32 2012 [VMM][I]: Generating deployment file:
> /srv/cloud/one/var/80/deployment.0
> Wed Sep  5 14:42:32 2012 [VMM][I]: ExitCode: 0
> Wed Sep  5 14:42:32 2012 [VMM][I]: Successfully execute network driver
> operation: pre.
> Wed Sep  5 14:42:33 2012 [VMM][I]: Command execution fail: cat << EOT |
> /var/tmp/one/vmm/xen/deploy /srv/cloud/one/var//datastores/0/80/deployment.0
> <host ip> 80 <host ip>
> Wed Sep  5 14:42:33 2012 [VMM][I]: Error: Errors were found at line 13 while
> processing /srv/cloud/one/var//datastores/0/80/deployment.0:
> Wed Sep  5 14:42:33 2012 [VMM][I]: <devices><hostdev
> mode="subsystem",type="pci",managed="no"><source><address
> domain="0x0000",bus="0x03",slot="0x00",function="0x0"/></source></hostdev></devices>
> Wed Sep  5 14:42:33 2012 [VMM][I]:
> Wed Sep  5 14:42:33 2012 [VMM][E]: Unable
> Wed Sep  5 14:42:33 2012 [VMM][I]: ExitCode: 1
> Wed Sep  5 14:42:33 2012 [VMM][I]: Failed to execute virtualization driver
> operation: deploy.
> Wed Sep  5 14:42:33 2012 [VMM][E]: Error deploying virtual machine: Unable
> Wed Sep  5 14:42:33 2012 [DiM][I]: New VM state is FAILED
> --------------------------------------
>
> The deployment file is:
> ----------------------------------
> name = 'one-80'
> #O CPU_CREDITS = 256
> memory  = '1024'
> bootloader = "hvmloader"
> disk = [
>     'file:/srv/cloud/one/var//datastores/0/80/disk.0,xvda,w',
>     'file:/srv/cloud/one/var//datastores/0/80/disk.1,xvdb,r',
> ]
> vif = [
>     ' mac=02:00:0a:01:20:0b,ip=<host ip>,bridge=br0',
> ]
> vfb = ['type=vnc,vnclisten=0.0.0.0,vncdisplay=80']
> <devices><hostdev mode="subsystem",type="pci",managed="no"><source><address
> domain="0x0000",bus="0x03",slot="0x00",function="0x0"/></source></hostdev></devices>
> -------------------------------------
>
> You can see that the error is at line 13. So is there a way for me to pass
> this through the RAW section or does anyone have an idea on how to do this?
>
> thanks and regards
> yasz
>
>
>
>
>
>
> _______________________________________________
> Users mailing list
> Users at lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>



-- 
Ruben S. Montero, PhD
Project co-Lead and Chief Architect
OpenNebula - The Open Source Solution for Data Center Virtualization
www.OpenNebula.org | rsmontero at opennebula.org | @OpenNebula