Osier,
Thank you very much for your helpful reply. I think I missed some info:
1. I execute nodeinfo, dumpxml command in windows client throw out of memory error, but can get right result in linux client, use the same VMware ESX 4.0 server, and virsh version is both 0.9.3
2. I want to try this with libvirt 0.9.10, but I don't where to download the missing 'libvirt-0.9.10-mingw.patch', could you tell me?
 
Thank you.
 
 
 

Mark
 
 
发件人: Osier Yang
发送时间: 2012-03-22 13:03
收件人: li_fuqiong
抄送: libvirt-users
主题: Re: [libvirt-users]转发: Error when executing virsh command to ESX
On 2012年03月21日 15:49, li_fuqiong wrote:
> To Whom It May concern:
> I found that the following error is caused by adding a scsi disk to
> install the os, I remove the scsi disk and the error remove but still
> report 'out of memory':
 
When trying to ask for help, it's better to tell which hypervisor
driver you are using.
 
> virsh # dumpxml win2003_122
> internal error Invalid or not yet handled value '/vmfs/devices/genscsi/mp
> x.vmhba0:C0:T1:L0' for VMX entry 'scsi0:1.fileName'
 
 From the source and per the error you got, we can known you are
using VMWare driver though.
 
VMWare driver parses the disk conf from .vmx file, and it recongnizes
the disk type with the disk file name suffix. Think you will fix
the problem by append ".vmdk" to the scsi disk you use in .vmx.
 
> virsh # list --all
> Id Name State
> ----------------------------------
> 80 win2003_122 running
> - winxp121 shut off
> virsh # nodeinfo
> error: failed to get node information
> error: out of memory
 
VMWare driver doesn't support nodeinfo, it's expected to see error
lile "this function is not supported by the connection driver". While
you see this, it's a bug. Could you try with libvirt 0.9.10 to see
if the problem is fixed yet?
 
Regards,
Osier