[Libvirt-cim] Re: KVM on Pegasus Test Run Summary for Sep 23 2008

[Deepti B Kalakeri]

Deepti B Kalakeri wrote:
> =================================================
> KVM on Pegasus Test Run Summary for Sep 23 2008
> =================================================
> Distro: Fedora release 9.90.1 (Rawhide)
> Kernel: 2.6.27-0.322.rc6.fc10.x86_64
> libvirt: 0.4.5
> Hypervisor: QEMU 0.9.1
> CIMOM: Pegasus 2.7.1
> Libvirt-cim revision: 688
> Libvirt-cim changeset: bc734ecd67ae
> =================================================
> FAIL      : 6
> XFAIL     : 3
> SKIP      : 6
> PASS      : 120
> -----------------
> Total     : 135
> =================================================
> FAIL Test Summary:
> ElementConforms - 01_forward.py: FAIL
> HostSystem - 01_enum.py: FAIL
> HostSystem - 03_hs_to_settdefcap.py: FAIL
The above test cases are failing because there is a mismatch in the 
hostname value returned by the providers and the hostname info used in 
the test cases.
The provider is returning the FQDN and the test case is just using the 
hostname.
Need to fix the above test cases.
The provider is returning an FQDN in case of RHEL5.2 and is independent 
of the order in which the information for 127.0.0.1 is present in 
/etc/hosts, but in case of F10 the information returned by the *ein* on 
HostSystem is completely dependent on the order in the /etc/hosts file. 
On F10 the first entry in the /etc/hosts is returned.
Should we use both the hostname and the FQDN to verify the HostSystem 
info in the test cases?
What is the expected behaviour from the providers ?
> LogicalDisk - 03_ld_gi_errs.py: FAIL
> Processor - 03_proc_gi_errs.py: FAIL
> SettingsDefine - 03_sds_fwd_errs.py: FAIL
The tc were failing bcs of the changes as part of the *"Add a little 
debugging to the device path"* provider patch.
The test cases were also failing because of the problem in the debug 
message in the common_util.py library.
Fixed and submitted the patches for the above three failures.

Thanks and Regards,
Deepti.