[Libvirt-cim] [PATCH 0 of 3] [TEST] Add ElementAllocatedFromPool for XenFV & KVM support

Kaitlin Rupert kaitlin at linux.vnet.ibm.com
Wed Apr 23 20:30:39 UTC 2008


Guo Lian Yun wrote:
> libvirt-cim-bounces at redhat.com wrote on 2008-04-21 23:31:45:
> 
>  > Guo Lian Yun wrote:
>  > > Signed-off-by: Guolian Yun <yunguol at cn.ibm.com>
>  >
>  > EAFP 01_forward.py and EAFP 02_reverse.py both fail for me on F9 with
>  > KVM and updated libvirt-cim sources.
>  
>   Could you please send the error log to me? Or do you know why these
>   are failing? It pass for me on F9 with KVM and recent libvrit-cim sources.

Here's the logs - I should have included them originally.

EAFP 01 is failing due to the disk pool issue.  You'll need to create a 
function that makes sure the guest's disk resource belongs to a pool. 
I'd suggested this in a previous mail, and I know you said you were 
going to add such a function.  But I haven't seen a patch for that yet - 
did I miss it?

Also, what version of libvirt are you using?  The disk pool support is 
new in 0.4.0.  I'd assume that F9 would have libvirt 0.4.0 or higher. 
It's possible your F8 system does not have a new version, which would 
cause this test to pass.

ElementAllocatedFromPool - 01_forward.py: FAIL
ERROR   - ElementAllocatedFromPool returned 0 ResourcePool 
                         objects for domain 'hd_domain'
CIM_ERR_FAILED: Unknown pool membership for `hd_domain/hda'

-----

ElementAllocatedFromPool - 02_reverse.py: FAIL
ERROR   - AttributeError : 'list' object has no attribute 'InstanceID'
CIM_ERR_NOT_FOUND: No such instance (foo)

This is because the InstanceID used for the diskpool is incorrect. This 
InstanceID will work on libvirt versions older than 0.4.0, but newer 
versions need to use the diskpool ID.

> 
>  >
>  > EAFP 01 and 02 are failing on Xen.  I'd like to see test cases fixed
>  > before we add KVM and XenFV support - otherwise, we can't test the KVM
>  > and XenFV support properly.
> 
>   They are fail for Xen on my own machine(F8), but pass on other machine
>   on F8, also pass on XenFV tested by Deepti. Of course, I will look
>   into this issue.
> 

I'm guessing the XenFV system probably has an older version of libvirt.

Sorry for all the confusion!  It's tricky making tests work for multiple 
guest types, as well as with differing package versions.

-- 
Kaitlin Rupert
IBM Linux Technology Center
kaitlin at linux.vnet.ibm.com




More information about the Libvirt-cim mailing list