[Libvirt-cim] [PATCH] [TEST] Update ResourceAllocationFromPool.01~04 for LXC support

[Kaitlin Rupert]

> [TEST] Update ResourceAllocationFromPool.01~04 for LXC support
> 
> diff -r 3ac66cf562f0 -r 7d2ce92ea89e suites/libvirt-cim/cimtest/ResourceAllocationFromPool/01_forward.py
> --- a/suites/libvirt-cim/cimtest/ResourceAllocationFromPool/01_forward.py	Fri May 30 00:24:45 2008 +0800
> +++ b/suites/libvirt-cim/cimtest/ResourceAllocationFromPool/01_forward.py	Fri May 30 14:22:08 2008 +0800
> @@ -31,7 +31,7 @@
>  from CimTest.Globals import logger, do_main
>  from CimTest.ReturnCodes import PASS, FAIL, XFAIL
> 
> -sup_types = ['Xen', 'XenFV', 'KVM']
> +sup_types = ['Xen', 'XenFV', 'KVM', 'LXC']
> 
>  @do_main(sup_types)
>  def main():
> diff -r 3ac66cf562f0 -r 7d2ce92ea89e suites/libvirt-cim/cimtest/ResourceAllocationFromPool/02_reverse.py
> --- a/suites/libvirt-cim/cimtest/ResourceAllocationFromPool/02_reverse.py	Fri May 30 00:24:45 2008 +0800
> +++ b/suites/libvirt-cim/cimtest/ResourceAllocationFromPool/02_reverse.py	Fri May 30 14:22:08 2008 +0800
> @@ -31,7 +31,7 @@
>  from CimTest.Globals import logger, do_main
>  from CimTest.ReturnCodes import PASS, FAIL, XFAIL
> 
> -sup_types = ['Xen', 'XenFV', 'KVM']
> +sup_types = ['Xen', 'XenFV', 'KVM', 'LXC']
> 
>  @do_main(sup_types)
>  def main():

Containers currently only really support memory and disk, so we need to 
make sure these only verify those two pools.

These tests don't currently support network and disk, but Deepti is 
working on adding that support.

Also, this test case passes, but it really shouldn't since there's no 
processor device returned from the container.  I think the issue is here 
(this is from ResourceAllocationFromPool/01_forward.py)

     for j in range(len(procdata)):
         if procdata[j].classname != get_typed_class(options.virt, 
"ProcResourceAllocationSettingData"):
             logger.error("ERROR: Association result error")
             status = FAIL

We need this test to instead, verify that the 
ProcResourceAllocationSettingData instance returned from the HRP 
association matches the one we expect for the guest.

This test isn't creating a guest though, and it really should be.

So, instead of adding LXC support to the test, I would fix these two 
test cases first so that they do the following:

1) Also support network and disk
2) Create a guest, and then verify the RASDs of the guest match the 
expected pools (and vice versa).

But check with Deepti, because I know that she is working on this 
already (see the "Fixing 02_reverse.py of RAPF" patch).
-- 
Kaitlin Rupert
IBM Linux Technology Center
kaitlin at linux.vnet.ibm.com