[Libvirt-cim] Test Run Summary (Jun 02 2009): KVM on Fedora release 10 (Cambridge) with sfcb

[Kaitlin Rupert]

> --------------------------------------------------------------------
> ComputerSystemIndication - 01_created_indication.py: FAIL
> ERROR 	- Exception : Request Failed: 200 
> Traceback (most recent call last):
>   File "./lib/XenKvmLib/const.py", line 139, in do_try
>   File "01_created_indication.py", line 146, in main
>     sub_list, ind_names, dict = sub_ind(ip, virt)
>   File "01_created_indication.py", line 60, in sub_ind
>     sub.subscribe(dict['default_url'], dict['default_auth'])
>   File "/data/users/kaitlin/sandbox/cimtest/suites/libvirt-cim/lib/XenKvmLib/indication_tester.py", line 345, in subscribe
>     "CreateInstance", auth_hdr)
>   File "/data/users/kaitlin/sandbox/cimtest/suites/libvirt-cim/lib/XenKvmLib/indication_tester.py", line 330, in __do_cimpost
>     (resp.status, resp.reason))
> Exception: Request Failed: 200 
> ERROR 	- None

I've seen issues with unsubscribing to indications on sfcb.  I think bit 
of code in the test case that does the unsubscribe needs to be modified. 
    If there are too many stale indication subscriptions hanging around, 
then there can be troubles adding a new subscription.

> --------------------------------------------------------------------
> HostSystem - 03_hs_to_settdefcap.py: FAIL
> ERROR 	- 'KVM_SettingsDefineCapabilities' returned 8 RASD objects instead of 4
> Class not found

The patch on the mailing list fixes this issue.

> --------------------------------------------------------------------
> RedirectionService - 01_enum_crs.py: FAIL
> ERROR 	- TypeError : __call__() takes exactly 1 argument (2 given)
> Traceback (most recent call last):
>   File "./lib/XenKvmLib/const.py", line 139, in do_try
>   File "01_enum_crs.py", line 108, in main
>     if res_val != exp_val:
> TypeError: __call__() takes exactly 1 argument (2 given)
> ERROR 	- None
> Class not found


The patch on the mailing list fixes this issue.

> --------------------------------------------------------------------
> ResourcePoolConfigurationService - 04_CreateChildResourcePool.py: FAIL
> ERROR 	- Exception in create_pool()
> ERROR 	- Exception details: (1, u'Pool with that name already exists')
> ERROR 	- Error in networkpool creation
> InvokeMethod(CreateChildResourcePool): Pool with that name already exists

Deepti is looking in to this issue.

> --------------------------------------------------------------------
> ResourcePoolConfigurationService - 07_DeleteResourcePool.py: FAIL
> ERROR 	- Exception in create_pool()
> ERROR 	- Exception details: (1, u'*** Provider Virt_ResourcePoolConfigurationService(2680) exiting due to a SIGSEGV signal ')
> ERROR 	- Error in networkpool creation
> InvokeMethod(CreateChildResourcePool): *** Provider Virt_ResourcePoolConfigurationService(2680) exiting due to a SIGSEGV signal 

The issue is that the test case is passing "nat" for the ForwardMode, 
but ForwardMode is a uint16 value.

Strangely, this issue doesn't appear on Pegasus, but sfcb seg faults 
when it tries to pull the CMPIObjectPath from the instance.  Will need 
to put a check in place, so sfcb doesnt segfault, and also fix the test 
case itself.

> --------------------------------------------------------------------
> SettingsDefineCapabilities - 01_forward.py: FAIL
> ERROR 	- KVM_SettingsDefineCapabilities returned 8 ResourcePool objects instead of 4

The patch on the mailing list fixes this issue.

> --------------------------------------------------------------------

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